JP5804107B2 - Method for producing silicon single crystal - Google Patents

Description

The present invention relates to a technique suitable for use in a method and apparatus for producing a silicon single crystal.
This application claims priority based on Japanese Patent Application No. 2009-145248 filed in Japan on June 18, 2009, and Japanese Patent Application No. 2009-14695 filed on June 19, 2009 in Japan. And the contents thereof are incorporated herein.

As a method for producing a silicon single crystal that is a material of a silicon wafer, a growing method by the Czochralski method (hereinafter, CZ method) is known.

  A method called CZ method is used for manufacturing a silicon single crystal. For example, in a vacuum inert gas (Ar) gas atmosphere, raw material polycrystalline silicon housed in a quartz crucible is melted by a heating means such as a resistance heater. A seed crystal (seed) is immersed in the surface of the silicon melt that is near the melting point after melting (seed melt contact process), and the liquid temperature is adjusted to such an extent that the seed crystal is compatible with the silicon melt. In order to remove dislocations, seed drawing with a diameter of about 5 mm is performed while pulling the seed upward (neck process). After the necking process for seed squeezing, the crystal diameter is expanded in a conical shape while adjusting the liquid temperature and the pulling speed so as to obtain the product diameter (shoulder process). When the crystal diameter reaches the product diameter, the product part is grown to a certain length in the vertical direction (straight barrel process), and then the crystal diameter is reduced to a conical shape (tail process). Cut off from the melt and finish.

  In the CZ method, when the silicon seed crystal is brought into contact with the raw material silicon melt, dislocations (thermal shock dislocations, misfit dislocations) occur in the seed crystal at high density due to thermal shock. This dislocation inhibits the single crystallization of the crystal to be grown. Therefore, it is necessary to proceed with subsequent crystal growth after eliminating the dislocation. In general, this dislocation is less likely to occur as the cross-sectional area of the growing crystal is smaller. In addition, the growth of dislocations once generated tends to stop. For this reason, in single crystal growth by the CZ method or the like, a method called a dash necking method in which a neck portion having a small cross-sectional area is grown to a predetermined length to eliminate dislocations is widely used.

  In the dash necking method, in the seed squeezing (necking) process in which the silicon seed crystal is melted by bringing it into contact with the raw material silicon melt, the neck portion once reduced to about 4 mm in diameter is about 50 to 200 mm in length. Form. Thereafter, a technique of thickening the single crystal by the shoulder portion until a predetermined diameter is obtained can be employed. As a result, dislocations propagated from slip dislocations introduced into the seed crystal can be eliminated in the neck portion, and a dislocation-free silicon single crystal can be pulled up.

Moreover, when pulling up a large single crystal, the introduction of dislocation during landing is suppressed so as not to give a heat shock. Techniques similar to these are described in Patent Documents 1 to 3.
Alternatively, as shown in Patent Document 5, there is a method in which a large diameter portion having an enlarged diameter is created in front of the shoulder shape, and the crystal is held by gripping the large diameter portion.

  In addition, Patent Document 4 describes a technique for demonstrating that there is no dislocation by growing only the neck portion in the necking step.

Japanese Patent Laid-Open No. 11-60379 JP 2008-87994 A Japanese Patent Laid-Open No. 9-2898 JP 2001-130996 A Japanese Patent Laid-Open No. 11-12082

SYNCHROTRON RADIATION INSTRUMENTATION: Ninth International Conference on Synchrotron Radiation Instrumentation.AIP Conference Proceedings, Volume 879, pp. 1545-1549 (2007).

  However, the techniques described in Patent Documents 1 to 3 have a problem that complete dislocation-freeness cannot be realized. In particular, even when the neck diameter is reduced to, for example, about 4 mm and grown to a length of about 500 mm, which is conventionally considered to be sufficient for dislocations, dislocations may remain in the straight body portion.

Further, as described in Patent Document 4, when the inside of the furnace is broken to take out the neck, contamination due to contaminants from the outside of the furnace or heavy metal contaminants from the moving part of the manufacturing apparatus occurs, resulting in a quartz crucible. It is difficult to reuse the raw material remaining inside. At present, when the diameter of the wafer is increased, it is impossible to waste a large amount of silicon raw material, so that it cannot be applied to actual manufacturing, and a method to replace it has been demanded.
Further, when a movable part as described in Patent Document 5 is provided in the furnace, the movable part becomes a contamination source, and as a result, there is a problem that it cannot be used for pulling a silicon single crystal.

  Furthermore, there is a problem that the state of occurrence and removal of dislocations cannot be accurately grasped.

  Furthermore, the mechanism of dislocation generation and removal is not accurately understood. Therefore, it has not been understood how dislocations can be completely removed. In particular, among the dislocations generated in the neck portion, there exist dislocations (axial dislocations) that extend in the crystal pulling direction and cannot be removed even if the crystal diameter is changed by a dash neck method or the like. These axial dislocations (dislocations that are difficult to remove) exist over the entire length of the straight body, and it was unclear how they could be removed.

The present invention has been made in view of the above circumstances, and achieves the following objects.
1. Dislocations generated in the seed melt contact process are surely removed to achieve dislocation-free.
2. Accurately capture the state of occurrence and removal of dislocations, especially axial dislocations.
3. To accurately grasp the relationship between dislocation behavior at the neck and the lifting conditions.
4). Accurately determine lifting conditions that can eliminate dislocations.

  In order to grow a large-diameter silicon crystal having a diameter of 300 mm or more, particularly about 450 mm, the diameter of the neck portion (dash neck portion) that supports a large weight becomes a problem. Therefore, the inventors of the present application have studied to grow safely by increasing the diameter of the portion conventionally referred to as the neck portion while maintaining the dislocation-free state. Here, it is assumed that a heavy silicon single crystal having an order of several hundred kilograms to several t is manufactured by a simple method without placing a special device as a contamination source in the furnace.

  In order to grow a large-diameter silicon single crystal, the above-described normal dash neck method is used. In the dash neck method, the diameter is reduced during the growth of a silicon single crystal, and dislocations introduced when the silicon seed crystal comes into contact with the silicon melt (heat shock dislocation, misfit dislocation) are removed. In this case, dislocations are more easily removed when the diameter of the neck portion is set to be thinner. Moreover, in order to grow a large weight crystal, it is necessary to set the diameter dimension to be thick from the load resistance of the neck portion. The thickness of the neck part was set to the dimension calculated from these relationships. In the case of a silicon single crystal having a diameter of 300 mm, the crystal load could be supported with a diameter of about 5 mm. However, in the case where the diameter of the silicon single crystal is 450 mm, when the crystal is grown to a necessary length, there is a possibility that the load cannot be supported by the neck portion.

  In addition, as described above, in order to make the neck portion thicker, there is a method in which the temperature difference between the surface of the seed and the melt (silicon melt) is reduced to prevent heat shock dislocation. there were. However, there have been cases where the silicon single crystal has been successfully grown without making the neck portion thinner or raising the seed temperature.

  Therefore, the inventors of the present application believe that if such conditions can be clarified, it is possible to grow large-sized silicon crystals by a simple method without the need for special treatments and apparatuses, and the behavior of dislocations inside the neck portion accompanying the pulling is considered. Sought a means to scrutinize.

  The inventors of the present application examined how dislocations behave in the necking process. We also examined why axial dislocations cannot be removed. Then, if the behavior of the dislocation could be clarified, it was thought that it was possible to grow a silicon crystal from which the dislocation was completely removed by a simple method without the need for special treatment or equipment. And we searched for a means to scrutinize the behavior of dislocations inside the neck part with the pulling.

  Conventionally, dislocation state analysis in the neck portion has been performed using X-ray topography. In this method, the cross section of the neck portion is sliced to a thickness of about 1.5 mm, and then the surface is etched with a mixed acid or the like to observe X-ray transmission. Therefore, since only an image of a certain cross section can be directly observed, the images at intervals of 1.5 mm are connected to capture the state change of dislocation accompanying crystal growth. Therefore, only dislocations propagating obliquely with respect to the pulling axis, which is the central axis in the pulling direction of the silicon single crystal, could be observed.

However, when white X-rays are used as high-energy radiation at the level described in Non-Patent Document 1, it has been found that dislocations disappear in a loop when the neck portion is observed. It was also found that dislocations exist in addition to the normal {111} plane.
That is, the present inventors have revealed the existence of the above-described axial dislocation for the first time by such a technique. Axial dislocations were observed as pits in the straight body of a silicon single crystal and a sliced wafer. However, it was not recognized that the observed pits were axial dislocations.

This axial dislocation can be clearly detected from the other pits on the wafer by the following method.
First, a silicon single crystal obtained by slicing a wafer has a crystal habit line properly formed and can realize dislocation free, and is recognized as a single crystal by other inspections. On such a wafer surface, axial dislocations can be identified by performing strain measurement using an optical inspection means. As such an inspection means, a strain inspection apparatus (SIRD; registered trademark) SirTec manufactured by JENAWAVE, which can visually observe the internal stress distribution state, can be used.

  At this time, the silicon wafer to be evaluated is subjected to a revealing treatment by heating at 900 to 1250 ° C. × 1 sec or more at a temperature rising / lowering rate of 10 to 300 ° C./sec. Thereafter, the surface properties of the wafer including the manifested strain are evaluated. This manifestation process is intended to generate a large temperature difference in the wafer surface by rapid heating and to generate thermal stress due to this temperature difference. Therefore, a long heating time is not required, and the desired effect can be exhibited even with a short heating time of about 1 sec.

Since the axial dislocation exists through the front and back surfaces of the wafer, as described above, distortion due to the thermal stress generated by the manifestation process, which is rapid heating for an extremely short time, increases. On the other hand, in surface contamination inspection using a laser surface inspection machine, dislocations that are detected as pits in the same way as axial dislocations but not long enough to penetrate the front and back surfaces of the wafers were generated by the manifestation process. It is considered that the strain due to thermal stress does not increase. That is, dislocations whose dimensions are less than about 10% of the thickness of the wafer are considered not to increase the strain due to the thermal stress generated by the revealing process.
For this reason, axial dislocations are manifested by the above-described manifestation process, but other pits and the like are not manifested by this manifestation process. FIG. 11 shows an observation example of a wafer having a manifested axial dislocation and an observation example of a wafer having no axial dislocation.
The observation example shown in FIG. 11A is an example in which three axial dislocations have occurred, and the observation example shown in FIG. 11B is an example in which no axial dislocation has occurred. In addition, it is thought that the distortion observed at the edge of the wafer is caused by the support of the wafer during the heat treatment.

  Axial dislocation occurs when the seed crystal is brought into contact with the melt. Axial dislocations are dislocations that continuously exist in the crystal growth direction up to a necking step, a shoulder portion formation (expansion) step, and a straight body portion growth step. Axial dislocations exist only in specific areas of the wafer when the wafer is manufactured. In addition, axial dislocations are removed depending on a conventionally used technique for growing a neck part, such as a diameter expansion, a diameter reduction, or a neck part length exceeding a certain value as in the dash neck method. I couldn't.

As shown in FIG. 12, the silicon single crystal 60 that has been pulled up has a seed crystal (seed crystal) T, a reduced diameter portion (neck portion) N0 as a silicon single crystal grown across the seeding interface T00, and an expanded diameter. A shoulder portion 60a, a straight body portion 60b, and a tail portion 60c.
When the vicinity of the seed crystal T is enlarged, as shown in FIG. 13, heat shock dislocations Jn and misfit dislocations Jm are generated in the reduced diameter portion (neck portion) N0. The heat shock dislocation Jn is generated on the seed crystal T side by the heat shock when the seed crystal T is brought into contact with the silicon melt, and is taken over to the neck portion N0 side of the growing silicon single crystal 60. The misfit dislocation Jm occurs when there is a mismatch in lattice constant between the seed crystal T and the neck portion N0 of the silicon single crystal 60.

  The axial dislocation J grows in the growth direction (axial direction) among the heat shock dislocation Jn and the misfit dislocation Jm. Specifically, as shown in FIG. 12, when viewing a 110 wafer W sliced from the single crystal 60, the axial dislocations J are 10 when the notch is in the 12 o'clock direction on the surface. Observed in the hour and 4 o'clock directions. More specifically, assuming that the notch in the 100 direction is 0 °, a region in the range of 120 ° to 135 °, and a region in the range of 315 ° to 350 ° symmetrical to the central axis of this region and the silicon single crystal 60. Observed at. Further, these regions are observed only in the range of the difficult-to-remove dislocation existence region J1 composed of the regions rotated by 90 ° and 45 ° about the central axis of the silicon single crystal 60. That is, the axial dislocation J extends in the growth direction (center axis direction) of the silicon single crystal 60 in the region J1. In the figure, the region J1 shows only regions in the range of 120 ° to 135 ° and 315 ° to 350 °.

  Note that in a silicon single crystal grown to a predetermined length or more, axial dislocations may disappear. However, at present, the manufacturing site recognizes that it is an axial dislocation only after combining the data of a wafer obtained by slicing a single silicon crystal that has been pulled up. Axial dislocations usually exist over the entire length of a single silicon single crystal and often penetrate the silicon single crystal.

  Here, it is preferable to use white X-ray topography by white X-rays as high energy radiation light for measuring dislocations in a silicon single crystal pulled by the CZ method. In particular, X-rays having energy of about 30 keV to 1 MeV, about 40 keV to 100 keV, and about 50 keV to 60 keV having a continuous spectrum are used. Alternatively, X-rays having a wavelength of about 0.001 nm to 0.25 nm are used. Further, X-rays having a distribution as shown in FIG. 5 are used in which the number of photons obtained through a 1 mm × 1 mm slit at a distance of 44 m from the light source. The graph shown in FIG. 5A shows the X-ray state as an example of the high-energy radiated light in the present invention, and shows the distribution of the number of photons with respect to the X-ray energy. The graph shown in FIG. 5B shows X-rays as an example of high-energy radiation light in the present invention, and shows the distribution of the number of photons with respect to wavelength. Further, the X-ray beam diameter is preferably 0.01 to 1 times the diameter of the neck portion which is the object to be measured.

  By using such high-energy white X-rays, the behavior of dislocations inside single crystal silicon can be observed as a three-dimensional image without performing a destructive inspection such as slicing, and information on the position of the dislocations. It has become possible for the first time to observe the nature of this.

  As shown in FIG. 4, paying attention to one diffraction spot (for example, 004 diffraction spot) and rotating the silicon single crystal, the CT based on the topographic image is reconstructed. Thereby, the three-dimensional position information of the dislocation line is obtained. At the same time, white X-ray topography as shown in FIG. 6 is performed to determine dislocation properties (such as Burgers vector).

X-ray topography is also called X-ray diffraction microscopy, and is a method for observing the spatial distribution of defects in a non-destructive manner. When X-rays are continuously incident on the silicon single crystal, a plurality of diffraction images called diffraction spots are observed. X-ray topography analyzes this diffraction spot.
Here, the following is mentioned as a two-dimensional detector used in order to observe a diffraction image. An imaging plate that utilizes the photoluminescent phenomenon of a fluorescent plate, X-ray film, nuclear dry plate, and photostimulable phosphor (BaFBr: Eu2 +). An X-ray television using an imaging tube having a PbO film sensitive to X-rays or an amorphous Se-As film as a photoconductive surface. CCD type X-ray detector using a charge coupled device CCD. As the X-ray detector used in the present invention, it is preferable to consider the following characteristics in terms of spatial resolution and dynamic range.

Detection quantum efficiency Dynamic range intensity linearity area dead time and counting light receiving area and position resolution Sensitivity nonuniformity Nonlinearity (or image distortion) of position
Energy resolution Time resolution Real-time measurement capability Operational stability

  In pulling a silicon single crystal, even if the same apparatus (CZ furnace) and the same in-furnace member such as carbon are used in the same manner, the state of dislocation generation and the behavior change depending on the pulling conditions. . Here, the pulling conditions include, for example, the pulling speed, the height (distance) from the surface of the silicon melt to the heat shielding member, and the seed crystal before contact with the silicon melt. It is time etc. to hold above. From this, using the means to observe the dislocation behavior inside the silicon single crystal by high energy synchrotron radiation, the neck part (dislocation) when using the manufacturing equipment adopting the structure combining different furnaces and different carbon materials It is possible to determine the manufacturing conditions of the removal section and the dislocation-free section.

  Therefore, by performing observation by X-ray topography, it is possible to exclude the production conditions in which axial dislocations are generated and specify the production conditions in which axial dislocations are not generated. The manufacturing conditions here are related to the characteristics of all single crystals and wafers, in addition to the so-called control parameter conditions, and include devices, means, parts, states, and processes for pulling and manufacturing single crystals.

  In order to grow a silicon single crystal having a large diameter of 300 mm or more, particularly about 450 mm, the diameter of the neck portion (dash neck portion) that supports a large weight becomes a problem. Accordingly, the inventors of the present application have used a simple method to mount a large-sized silicon single crystal having an order of several hundred kilograms to several t without axial dislocations, without placing a special device as a contamination source in the furnace. For the purpose of nurturing safely in a non-existent state, we examined increasing the diameter of the part that was previously called the neck part.

  In order to grow a large-diameter silicon single crystal, the above-mentioned dash neck method is usually used, and the neck portion is formed by reducing the diameter of the silicon single crystal at the initial stage of growth. Thereby, dislocations (heat shock dislocations, misfit dislocations) introduced when the silicon seed crystal comes into contact with the silicon melt are removed at the neck portion. In this case, dislocations are more easily removed when the neck portion has a smaller diameter. However, in order to grow a large weight crystal, it is necessary to increase the diameter dimension from the load resistance of the neck portion. The thickness of the neck portion was calculated from these relationships. A crystal having a diameter of about 300 mm can support the load of the crystal with a diameter of about 5 mm. However, if the crystal has a diameter of 450 mm, the load may not be supported if the required crystal length is set. Therefore, the inventors of the present application examined increasing the diameter of the neck as described above.

  The inventors of the present application have found that the above-described axial dislocation J may exist when the diameter of the neck portion is increased. Conventionally, such an axial dislocation has not been confirmed.

It has been found that the behavior of the heat shock dislocation Jn and the misfit dislocation Jm has the following characteristics. Here, the behavior of the heat shock dislocation Jn and the misfit dislocation Jm refers to the propagation direction in the process of growing a silicon single crystal that pulls up the neck portion, the shoulder portion, and the straight body portion.
1) As a major premise, dislocations move on the slip plane ({111} plane in the case of silicon).
2) Usually, dislocations are introduced at {111}.
3) In rare cases, other planes may be moved without being introduced at {111}. Also in that case, it is usually immediately changed to {111}.
4) Most of the dislocations introduced into the silicon single crystal produced by the CZ method occur when the seed crystal is brought into contact with the silicon melt.

  Here, Table 1 shows angles between crystal planes as slip planes where dislocations move and types of crystals to be pulled up.

  When a silicon single crystal for manufacturing a <100> wafer is manufactured by pulling it out of a silicon melt, the dislocation generated by heat shock or misfit is a necking step for reducing the diameter of the silicon single crystal as shown in FIG. In {circle over (1)}, {111}, which is the sliding surface of the silicon single crystal, is moved. Dislocations that have moved on the slip plane reach the surface of the silicon single crystal. The surface of the silicon single crystal is considered to be the end point of dislocation. However, according to the above X-ray topography, it was found that the initial dislocations introduced by heat shock or misfit may extend in the crystal growth direction. This phenomenon occurs regardless of the diameter of the product and the diameter of the neck portion during necking. At this time, the dislocation does not move {111} on the slip plane, extends in a direction perpendicular to the crystal growth interface between the silicon single crystal and the silicon melt, and then moves to the slip plane without moving to the slip plane. Progress in the growth direction. Dislocations that have progressed in the crystal growth direction in this way become axial dislocations.

The axial dislocation does not reach the surface of the silicon single crystal even if the diameter of the silicon single crystal is subsequently increased to the diameter of the product. Such a silicon single crystal having axial dislocations has a habit line similar to a silicon single crystal having no axial dislocations, and may not be distinguished from non-defective products in an ingot state. When the wafer sliced from the ingot is inspected, the axial dislocation is detected as a defect.
Also, dislocations extending in the growth direction of the silicon single crystal increase the diameter of the silicon single crystal, or the diameter reaches the product diameter, and the product part moves to {111} during manufacturing. In some cases, axial dislocations may occur.

  If the dislocation introduced by heat shock or misfit and extended parallel to the crystal growth direction can be moved to {111} during the necking step, the dislocation can be removed at the neck portion. That is, even if the neck portion is not made thinner than necessary or the seed is not heated to an unnecessarily high temperature, it is possible to proceed to a shoulder process in which the diameter of the silicon single crystal is increased with dislocations removed in the neck portion. . This eliminates the influence of dislocations generated in the initial stage of crystal growth in the product. As a result, the number and time of remelting (remelting) the silicon single crystal in which axial dislocations are generated in the initial stage of crystal growth can be suppressed.

  Therefore, the inventors of the present invention have considered that if such conditions can be clarified, it is possible to grow a silicon crystal having a large weight without axial dislocation at low cost, and the present invention has been completed.

The inventors of the present application focused on the following points.
1) A silicon single crystal grows perpendicularly to a crystal growth surface which is an interface between the silicon single crystal and the silicon melt. For this reason, dislocations that are generated and introduced when the neck portion is formed are not {111} but may move in a direction perpendicular to the crystal growth surface.
2) Accordingly, the present invention is a method for moving dislocations that do not exist in {111} to {111} at the stage of forming the neck portion. This method employs means for changing the shape of the interface between the silicon single crystal and the silicon melt when the neck portion is formed. X-ray topography in synchrotron radiation equipment is used for checking the conditions.

The method for producing a silicon single crystal of the present invention is a method for producing a silicon single crystal in which the silicon single crystal is pulled by a CZ method.
A dipping step of bringing a seed crystal into contact with the silicon melt and starting to pull up the silicon single crystal;
Dislocations generated in the dipping process are removed by extending the silicon single crystal in the outward direction according to the pulling up of the silicon single crystal or by annihilating it in a loop form, and dislocation removal or non-dislocation parts in the silicon single crystal Dislocation removal step to form
A diameter expanding step of pulling up the silicon single crystal so as to form a shoulder portion by expanding the diameter of the silicon single crystal to a required diameter;
A straight cylinder step of pulling up the silicon single crystal so as to form a straight cylinder portion having the required diameter in the silicon single crystal;
Dislocation behavior information acquisition step of detecting dislocation behavior information which is a three-dimensional behavior of the dislocation in the dislocation removal unit or the non-dislocation unit by irradiating the dislocation removal unit or the non-dislocation unit with high energy radiation. When,
Based on the dislocation behavior information, a dislocation-free determination step for determining that the dislocation has been removed,
Based on the determination of dislocation-free in the dislocation-free determination step, a pulling condition setting step for determining pulling conditions for dislocation-free the silicon single crystal;
Have
The above-described problem has been solved by irradiating the high-energy radiation light in the dislocation behavior information acquisition step with an energy within a range of 40 keV to 70 keV.
The method for producing a silicon single crystal of the present invention is a method for producing a silicon single crystal in which the silicon single crystal is pulled by a CZ method.
A dipping step of bringing a seed crystal into contact with the silicon melt and starting to pull up the silicon single crystal;
A straight body step of pulling up the silicon single crystal so as to form a straight body portion of the silicon single crystal to be a silicon wafer;
A dislocation removal or dislocation removal step that removes dislocations or difficult dislocations generated in the dip step before the straight body step; and
A diameter expanding step of pulling up the silicon single crystal so as to form a shoulder portion by expanding the diameter of the silicon single crystal to the diameter of the straight body portion after the difficult removal dislocation removing step;
Dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of the dislocation in the dislocation removal unit or the non-dislocation unit by irradiating the dislocation removal unit or the non-dislocation unit with high energy radiation;
Based on the dislocation behavior information, a dislocation-free determination step for determining that the dislocation has been removed,
Based on the determination of dislocation-free in the dislocation-free determination step, a pulling condition setting step for determining pulling conditions for dislocation-free the silicon single crystal;
And having
The high energy synchrotron radiation in the dislocation behavior information acquisition step is irradiated with energy in the range of 40 keV to 70 keV,
In the difficult-to-remove dislocation removal step, the dislocation generated in the dip step is removed by extending the silicon single crystal toward the outside according to the pulling up of the silicon single crystal or by annihilating it in a loop shape to form the silicon single crystal. The above problem has been solved by forming a dislocation removal part or a dislocation-free part.
The method for producing a silicon single crystal of the present invention is a method for producing a silicon single crystal in which the silicon single crystal is pulled by a CZ method.
A dipping step of bringing a seed crystal into contact with the silicon melt and starting to pull up the silicon single crystal;
The dislocations generated in the dip process are removed by extending the silicon single crystal toward the outside according to the pulling of the silicon single crystal or by annihilating it in a loop shape, and dislocation removal or non-dislocation parts are formed in the silicon single crystal. Dislocation removal step to be formed;
A diameter expansion step of pulling up the silicon single crystal so as to form a shoulder portion by expanding the diameter of the silicon single crystal to a required diameter;
A straight cylinder step of pulling up the silicon single crystal so as to form a straight cylinder portion having the required diameter in the silicon single crystal;
Dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of the dislocation in the dislocation removal unit or the non-dislocation unit by irradiating the dislocation removal unit or the non-dislocation unit with high energy radiation;
A non-dislocation determination step for determining that the dislocation has been removed based on the dislocation behavior information,
The high energy synchrotron radiation in the dislocation behavior information acquisition step is irradiated with energy in the range of 40 keV to 70 keV,
Based on the determination of non-dislocation in the non-dislocation determination step, the above-described problem is solved by starting the diameter expansion step.
The method for producing a silicon single crystal according to the present invention includes the following steps in the method for producing a silicon single crystal in which the silicon single crystal is pulled by the CZ method. A dipping process in which the seed crystal is brought into contact with the silicon melt to start pulling. A straight body process that pulls up the single crystal straight body that becomes a silicon wafer. A difficult-to-remove dislocation removal step of removing difficult-to-remove dislocations generated in the dip step before the straight body step. By having these steps, the above problems were solved.
In the difficult removal dislocation removal step, the extension direction of the difficult removal dislocation may be moved from the crystal growth direction.
In the difficult-to-remove dislocation removal step, the difficult-to-remove dislocations may be removed by extending outwardly according to crystal pulling or by annihilating them in a loop shape.
In the difficult-to-remove dislocation removal step, the solid-liquid interface shape between the silicon melt and the silicon single crystal may be deformed from the melt surface to move the extension direction of the difficult-to-remove dislocation from the crystal growth direction.
In the difficult-to-removal dislocation removal step, the extension direction of the difficult-to-removal dislocations is changed from the crystal growth direction by changing the portion where the tangential plane of the solid-liquid interface between the silicon melt and the silicon single crystal is parallel to the melt surface It may be moved.
In the difficult-to-removal dislocation removal step, fluctuation of the solid-liquid interface state between the silicon melt and the silicon single crystal may be increased to move the extension direction of the difficult-to-remove dislocation from the crystal growth direction.
In the difficult-to-remove dislocation removal step, the rotation of the silicon single crystal relative to the silicon melt may be reduced to increase the fluctuation of the solid-liquid interface state.
In the difficult-to-remove dislocation removal step, the magnetic field strength applied to the silicon melt may be reduced to increase the fluctuation of the solid-liquid interface state.
In the difficult-to-remove dislocation removal step, fluctuations in the solid-liquid interface state may be increased by molten-metal surface vibrations generated due to the quartz glass crucible storing the silicon melt.
In the difficult-to-remove dislocation removal step, the fluctuation of the solid-liquid interface state may be increased by heating the silicon single crystal asymmetrically with respect to the growth axis.
In the difficult removal dislocation removal step, a temperature distribution is formed so that the solid-liquid interface between the silicon melt and the silicon single crystal is asymmetric with respect to the growth axis, and the extension direction of the difficult removal dislocation is moved from the crystal growth direction. May be.
In the difficult-to-remove dislocation removal step, the silicon single crystal may be heated asymmetrically with respect to the growth axis so that the solid-liquid interface is asymmetric with respect to the growth axis.
The method for producing a silicon single crystal according to the present invention forms a dislocation removing portion that removes dislocations generated in the dipping step by extending the dislocations generated in the dip step in the outward direction or by annihilating them in a loop shape in the difficult removal dislocation removing step. In addition, the following steps are included. A diameter expansion step of pulling up a shoulder portion that expands to a necessary diameter of the straight body after the difficult removal dislocation removal step. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. A lifting condition setting step of determining a dislocation-free implementation raising condition based on the determination of dislocation-free in the non-dislocation determination step.
The method for producing a silicon single crystal according to the present invention forms a dislocation removing portion that removes dislocations generated in the dipping step by extending the dislocations generated in the dip step in the outward direction or by annihilating them in a loop shape in the difficult removal dislocation removing step. In addition, the following steps are included. A diameter expansion step of pulling up a shoulder portion that expands to a necessary diameter of the straight body after the difficult removal dislocation removal step. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. And based on the determination of non-dislocation in the non-dislocation determination step, the diameter expansion step is started.
In the above-described silicon single crystal manufacturing method, the high-energy radiation light may have an energy of 40 keV to 70 keV, and may be irradiated to the dislocation removing unit as a rotation state at a rotation speed of 0.1 to 30 rpm .

  The silicon single crystal manufacturing method of the present invention is a silicon single crystal manufacturing method in which the silicon single crystal is pulled up by the CZ method, and includes the dipping step, the straight body step, and the difficult-to-removal dislocation removal step as described above. This makes it possible for the first time to remove difficult dislocations (axial dislocations) that could not be removed in the past.

  In the difficult-to-remove dislocation removal step, the axial dislocation can be removed before the diameter expansion step of pulling up the shoulder portion by moving the extension direction of the axial dislocation from the crystal growth direction. Therefore, it becomes possible to pull up a silicon single crystal in which no dislocation extending to the straight body portion exists as an axial dislocation. Therefore, it becomes possible to manufacture a wafer having no axial dislocation from the straight body portion of the silicon single crystal without axial dislocation.

  In the difficult-to-removal dislocation removing step, the dislocations are removed by extending outwardly according to crystal pulling or by annihilating them in a loop shape. That is, the dislocation is moved to a slip plane that is different from the crystal growth direction. Then, the dislocations can escape to the outside or be paired with other dislocations. Thereby, generation | occurrence | production of an axial dislocation can be prevented. For this reason, it is possible to pull up a single crystal free from dislocations extending to the straight body as axial dislocations from the silicon melt. In this way, it was conventionally judged that it was single-crystallized, but in practice, incomplete crystals in which axial dislocations are present can be eliminated, and silicon single crystals without axial dislocations can be pulled up. It becomes possible.

In the difficult-to-remove dislocation removal step, the solid-liquid interface shape between the silicon melt and the silicon single crystal can be deformed from the melt surface to move the extension direction of the difficult-to-remove dislocation from the crystal growth direction.
The dislocation extension is not {111}, which is a slip plane, but first moves in that direction because the crystal grows perpendicular to the crystal growth plane. That is, as shown in FIG. 16A, the solid-liquid interface K between the silicon melt 3 and the grown single crystal 6 extends in the normal direction of the tangential plane with the dislocation j as a contact. When the solid-liquid interface K is parallel to the liquid surface 30, that is, when the solid-liquid interface K is in a flat state, the dislocation j extends in the direction indicated by the arrow in the melt 3 as shown in FIG. Go. Accordingly, the dislocation j in the axial direction (crystal growth direction) shown in the vertical direction in the drawing extends in the crystal growth direction. When the dislocation j extends in the crystal growth direction as it is, the position of the dislocation j changes in the radial direction at the enlarged diameter portion of the silicon single crystal, but is not erased. Therefore, the dislocation j extending in the crystal growth direction may be an axial dislocation that exists in the straight body portion so as to extend in the crystal growth direction.

The deformation of the solid-liquid interface shape between the silicon melt and the silicon single crystal from the surface of the melt means that the solid state between the silicon melt 3 and the grown silicon single crystal 6 is shown in FIG. This includes making the liquid interface K1 convex downward. Further, as shown in FIG. 16C, the solid-liquid interface K2 between the silicon melt 3 and the grown silicon single crystal 6 is formed in a convex shape upward.
As a result, the dislocation j extends in the normal direction of the tangential plane with the dislocation j as a contact at the solid-liquid interface K1, as shown in FIG. Further, as shown in FIG. 16C, the dislocation j extends in the normal direction of the tangential plane with the dislocation j as a contact at the solid-liquid interface K2 as the crystal is pulled.
Therefore, the dislocations j extend toward the outside of the silicon single crystal as shown in FIG. Alternatively, as shown in FIG. 16C, a plurality of dislocations j gather in the direction of the central axis of the silicon single crystal. As a result, the dislocation j can be removed outside the silicon single crystal, or can be eliminated by annihilation with other dislocations. Therefore, the dislocation j can be prevented from becoming an axial dislocation, that is, a dislocation that is difficult to remove.

The relationship between the solid-liquid interface shape and the direction of dislocation propagation can be summarized as follows.
(A) When the interface shape is flat, the dislocation extends in the vertical direction (the central axis direction of the silicon single crystal, the vertical direction). (B) When the interface shape protrudes downward, the dislocation extends toward the outer periphery of the silicon single crystal ( c) When the interface shape is convex upward, the dislocation extends toward the inside of the silicon single crystal.

These solid-liquid interface shapes can be controlled by the pulling speed. That is, when the pulling speed is increased, the shape of the interface becomes convex upward. Further, when the pulling speed is reduced, the shape of the interface becomes convex downward. In addition, more precise control can be performed by changing the magnitude of the magnetic field to be applied, the rotational speed of the crucible, the rotational speed of the crystal, and the like.
In addition, the diameter of the crystal may vary due to a change in the pulling speed. When the pulling rate is reduced in order to remove dislocations and prevent the occurrence of axial dislocations, the dislocations can be removed in the outward direction of the crystal, and the occurrence of axial dislocations can be prevented. In this case, there is no problem that the diameter is reduced to a certain extent so that the crystal to be pulled can be held. Further, when the pulling speed is increased, the dislocations can be eliminated by disappearing toward the center, or they can be removed as they are, so that the occurrence of axial dislocations can be prevented. In this case, there is no problem in expanding the diameter so as not to waste the raw material. Therefore, the range of the crystal diameter variation is not particularly limited as long as it does not affect the pulling up of the straight body of the silicon single crystal.
In order to remove dislocations, it is also preferable that the dislocation extension direction shifts from the pulling direction (axial direction) of the silicon single crystal due to a change in pulling speed.

  In the difficult-to-removal dislocation removal step, the extension direction of the difficult-to-removal dislocations is changed from the crystal growth direction by changing from the central position the portion where the tangential plane of the solid-liquid interface between the silicon melt and the silicon single crystal is parallel to the melt surface. Can be moved. Both the solid-liquid interface K1 having a downwardly convex shape shown in FIG. 16B and the solid-liquid interface K2 having an upwardly convex shape shown in FIG. 16C rotate around the crystal center. It is a rotationally symmetric shape with an axis. Accordingly, the shape is axisymmetric with respect to the axis of the central axis of the silicon single crystal. Dislocations may develop in line with the central axis of the silicon single crystal. That is, dislocations may exist at the crystal center. In this case, as described above, no matter how much the degree of the convex shape of the solid-liquid interface shape is changed, it is considered that the dislocation at the center of the crystal does not change the extension direction from the central axis direction.

  Changing the portion where the tangent plane of the solid-liquid interface between the silicon melt and the silicon single crystal is parallel to the melt surface from the center position includes the following. As shown in FIG. 17A, the solid-liquid interface K3 between the silicon melt 3 and the grown single crystal 6 is convex downward. In addition, the shape of the solid-liquid interface K3 is changed so as not to be axially symmetric so that the portion Kd in which the normal line of the tangential plane is the vertical direction, that is, the crystal center axis direction, is separated from the crystal center position Kc. . Alternatively, as shown in FIG. 17B, the solid-liquid interface K4 between the silicon melt 3 and the grown single crystal 6 is convex upward. In addition, the solid-liquid interface K3 shape is changed so as not to be axially symmetric so that the portion Kd in which the normal line of the tangential plane is vertical, that is, the crystal central axis direction is separated from the crystal center position Kc. .

  Control for shifting the shape of the solid-liquid interface K3 shown in FIG. 17A or the solid-liquid interface K4 shown in FIG. 17B from axial symmetry includes the following operations. The rotation ratio between the silicon melt 3 and the grown crystal 6 is made zero, that is, the crystal rotation with respect to the melt is stopped. Stop the applied magnetic field during pulling. The temperature distribution of the grown crystal is distorted so that it is not symmetric with respect to the central axis of the crystal. Changing the state of the silicon melt from a state suitable for crystal growth symmetrical to the central axis of the crystal.

  By reducing or stopping the rotation of the crystal relative to the melt, fluctuation occurs from the state in which the convection distribution of the silicon melt is axisymmetric with respect to the central axis of the crystal. Then, the symmetry of the convection distribution is broken and the crystal growth is not axisymmetric. As a result, for example, a silicon single crystal does not have a uniform diameter, but the crystal grows. That is, the temperature distribution which is axisymmetric with respect to the central axis of the silicon single crystal is broken at the solid-liquid interface. Then, the portion where the normal line of the tangent plane of the solid-liquid interface is in the vertical direction, that is, the central axis direction of the crystal can be moved to a position away from the center of the crystal.

  By stopping the magnetic field applied during the pulling, fluctuation occurs from the state in which the convection in the silicon melt is controlled by the magnetic field. The crystal growth is not axisymmetric with respect to the central axis of the crystal. As a result, crystals do not have a uniform diameter but grow. That is, the temperature distribution which is axisymmetric with respect to the central axis of the silicon single crystal is broken at the solid-liquid interface. Then, the portion where the normal line of the tangent plane of the solid-liquid interface is in the vertical direction, that is, the central axis direction of the crystal can be moved to a position away from the center of the crystal.

  In order to form a temperature distribution that is not axially symmetric in the crystal, only a biased portion (diameter portion) of the grown crystal is heated and cooled, or heated or cooled. Specifically, only a specific part near the solid-liquid interface is irradiated with laser light on the rotating crystal, and the part is heated. Further, in a furnace for pulling up the silicon single crystal, a gas flow is supplied in the vicinity of the surface of the silicon melt in the diameter direction of the crystal so that only one side of the crystal is cooled. The temperature control means for heating or cooling a part of the crystal can employ the following configuration. A structure in which the temperature adjusting means rotates following the rotation of the crystal. A configuration that can be switched on and off following the rotation of the crystal. A configuration that can be adjusted by rotating the position where the laser is irradiated or the position where the cooling gas is ejected. Such temperature adjusting means may be installed at a fixed position.

  The temperature adjusting means may form a temperature distribution that is not axially symmetric in the crystal so as not to shield the silicon melt from a part of the crystal. For example, a notch that is provided in a furnace for pulling up a silicon single crystal and that rotates in synchronization with the crystal may be provided near the lower end of a heat shield that reduces heat radiated from the silicon melt to the crystal. Specifically, a lifting device provided with a notch that can be opened and closed at the lower end of the heat shield and provided with a rotating means for rotating the heat shield in synchronization with the rotation of the crystal may be used.

In order to change the state of the silicon melt from a state suitable for crystal growth that is symmetric with respect to the central axis of the silicon single crystal, it is necessary to shift the heating of the quartz glass crucible storing the silicon melt from the symmetric state. Conceivable. For example, the heating position may be rotated following the rotating crucible.
In addition, a part of the quartz glass crucible, for example, a state in the range of about a quarter in the circumferential direction of the inner wall, is different from the other parts, and bumping that is considered to be the cause of the vibration of the silicon melt surface (liquid level) The occurrence rate may be increased. Specifically, the bubble content at a depth (thickness position) of 0.5 mm to 1 mm from the surface of the inner wall within a range of 10 cm from the upper end of the crucible and within a range of about ¼ of the circumferential direction. It may be increased by about 30% (25 to 35%) from the bubble content on the surface of the inner wall in the range other than.

  In addition to the above, the shape (state) of the solid-liquid interface can be changed by changing parameters such as pulling speed, applied magnetic field strength, crucible rotation speed, crystal rotation speed, crystal and silicon melt heating state, etc. Any method can be employed as long as the method is shifted from axial symmetry.

  In the difficult-to-removal dislocation removal step, fluctuation of the solid-liquid interface state between the silicon melt and the silicon single crystal may be increased to move the extension direction of the difficult-to-remove dislocation from the crystal growth direction. As described above, this fluctuation of the solid-liquid interface acts to shift the dislocation growth direction from the crystal growth direction by various means. Then, dislocations can be eliminated from the crystal so as to disappear, or can be eliminated from each other, thereby preventing the formation of axial dislocations.

  In the difficult-to-remove dislocation removal step, the rotation of the silicon single crystal relative to the silicon melt may be reduced to increase the fluctuation of the solid-liquid interface state.

  In the difficult-to-remove dislocation removal step, the magnetic field strength applied to the silicon melt may be reduced to increase the fluctuation of the solid-liquid interface state.

  In the difficult-to-remove dislocation removal step, fluctuations in the solid-liquid interface state may be increased by molten-metal surface vibration generated due to the quartz glass crucible storing the silicon melt.

  In the difficult to remove dislocation removal step, the fluctuation of the solid-liquid interface state may be increased by heating the silicon single crystal asymmetrically with respect to the growth axis.

  In the difficult removal dislocation removal step, a temperature distribution is formed so that the solid-liquid interface between the silicon melt and the silicon single crystal is asymmetric with respect to the growth axis, and the extension direction of the difficult removal dislocation is moved from the crystal growth direction. May be.

  In the difficult-to-remove dislocation removal step, the silicon single crystal may be heated asymmetrically with respect to the growth axis so that the solid-liquid interface is asymmetric with respect to the growth axis.

The method for producing a silicon single crystal according to the present invention forms a dislocation removing portion that removes dislocations generated in the dipping step by extending the dislocations generated in the dip step in the outward direction or by annihilating them in a loop shape in the difficult removal dislocation removing step. In addition, the following steps are included.
A diameter expansion step of pulling up a shoulder portion that expands to a necessary diameter of the straight body after the difficult removal dislocation removal step. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. A lifting condition setting step of determining a dislocation-free implementation raising condition based on the determination of dislocation-free in the non-dislocation determination step.
This makes it possible to clarify the behavior of dislocations whose behavior has not been clarified. Further, this dislocation can be eliminated so as to escape outward along with the growth of the dislocation removal portion (dislocation-free portion). Alternatively, it is possible to easily obtain a pulling condition for expanding the diameter of the silicon single crystal after eliminating the dislocation as a loop and eliminating the dislocation.

  At this time, the dislocation state is determined from the acquired information in the non-dislocation determination step. Therefore, it is possible to accurately grasp the dislocation behavior in the dislocation removal unit (dislocation-free unit). As a result, dislocation removal due to the occurrence of a loop, which has not been known in the past, can be used effectively. For this reason, it is not necessary to make the diameter of the dislocation-free portion, conventionally called the neck portion, smaller than necessary. Thereby, even if the weight of the crystal pulled up from the silicon melt increases, the load can be supported by the dislocation-free portion. Therefore, a silicon single crystal having a larger diameter than before can be pulled out of the silicon melt. Further, conventionally, extra steps related to the dipping process and the like have caused a long manufacturing time and an increase in work processes. However, according to the above-described method, it is possible to pull up a silicon single crystal having a large weight of about 450 mm in diameter without dislocation without axial dislocation without increasing the manufacturing time and increasing the number of work steps. It becomes possible.

The method for producing a silicon single crystal according to the present invention forms a dislocation removing portion that removes dislocations generated in the dipping step by extending the dislocations generated in the dip step in the outward direction or by annihilating them in a loop shape in the difficult removal dislocation removing step. In addition, the following steps are included. A diameter expansion step of pulling up a shoulder portion that expands to a necessary diameter of the straight body after the difficult removal dislocation removal step. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. And based on the determination of non-dislocation in the non-dislocation determination step, the diameter expansion step is started.
As a result, it is possible to observe in real time the behavior of dislocations generated during pulling up by non-destructive inspection, which has been impossible in the past. Therefore, after confirming that the dislocations are removed (dislocation-free) in the dislocation removal portion (dislocation-free portion), the shoulder portion (expanded diameter portion) and the straight body portion can be pulled up. Therefore, dislocations are removed, and dislocation-free crystals having no axial dislocations can be pulled up safely and easily.

In the silicon single crystal manufacturing method, the high-energy radiation light may have an energy of 40 keV to 70 keV.
With the high-energy synchrotron radiation having such energy, it is possible to sufficiently acquire information on the dislocation state necessary for growing a dislocation-free crystal.
Further, the dislocation removing unit may be irradiated with the high energy radiation light as a rotation state at a rotation speed of 0.1 to 30 rpm.
Thereby, the three-dimensional information of dislocation can be acquired. In addition, information can be acquired nondestructively.
Here, the irradiation with high energy radiation light in a rotating state includes the following. A dislocation removal part (dislocation-free part) rotates around the central axis in the crystal growth direction as a rotation center axis, and this dislocation-free part is irradiated with high energy radiation. Alternatively, irradiation with high energy synchrotron radiation for obtaining information on dislocation is performed in a relative motion state in which the irradiation source of high energy synchrotron radiation and the dislocation-free part are comparable to the above rotation.

  In the initial stage of pulling the silicon single crystal by the CZ method, a dipping process, a neck process, and a shoulder forming process are performed. In the dipping process, the seed crystal is lowered while rotating, and its tip is immersed in the surface of the silicon melt. After dipping the tip of the seed crystal in the silicon melt, the descent of the seed crystal is stopped. Then, the silicon melt and the seed crystal are sufficiently blended.

Usually, when the seed crystal is brought into contact with the silicon melt, a meniscus is formed at the interface due to the surface tension of the silicon melt. However, the temperature of the melt immediately after melting the crystal raw material has a large local temperature fluctuation. Further, the temperature of the melt as a whole is extremely large and unstable.
For this reason, the dipping process is performed after a predetermined time has elapsed after melting the crystal raw material. At this time, if the temperature of the liquid surface when the seed crystal is brought into contact with the silicon melt is too high, the tip of the seed crystal melts and is separated from the melt. Conversely, if the temperature of the silicon melt surface is too low, crystals grow around the tip of the seed crystal. Then, the crystal protrudes around the seed crystal on the surface of the melt. When shifting from the dip process to the neck process in such a state, new dislocations are generated in the neck part.

For this reason, when shifting from a dipping process to a neck process, it is necessary for the melt temperature to be stabilized. Therefore, after immersing the tip of the seed crystal in the silicon melt, fully blending the seed crystal and the silicon melt, and confirming that the temperature of the silicon melt has stabilized, the dipping process is followed by the neck process. Transition.
That is, conforming the seed crystal and the silicon melt includes the following. Estimate the temperature of the surface (liquid level) of the silicon melt by observing the shape of these interfaces when the seed crystal is brought into contact with the silicon melt. Control the heater power based on the estimated liquid surface temperature to adjust the heat input to the silicon melt.
In other words, conforming the seed crystal and the silicon melt includes the following operations. An operation for adjusting and stabilizing the melt surface temperature by adjusting the heater power so that a meniscus having a predetermined shape is formed around the tip of the seed crystal when the growth rate of the seed crystal is 0 (zero). . The meniscus is not formed only when the seed crystal and the silicon melt are mixed. The meniscus is also formed at the interface between the crystal and the melt during the growth of the crystal, such as a neck process following the process of allowing the seed crystal and the silicon melt to blend.

  Regardless of the CZ method or the MCZ method, the seed crystal is adapted to the melt in the dipping process, the neck part having a predetermined length is formed in the neck process, and then the formed neck part is adapted to the melt again. Form the neck. Thereby, the length of the neck part required for non-dislocation to remove the dislocation from the neck part can be shortened, and the non-dislocation rate can be improved.

(1) In the case of using the CZ method, first, the raw material of the crystal in the crucible is melted, and the seed crystal is immersed in a melt held in the crucible to make the seed crystal conform. Then, the neck process which raises a seed crystal and forms a neck part is implemented. Next, a shoulder portion and a body portion of a single crystal are formed. In such a silicon single crystal manufacturing method, after the neck portion having a predetermined length is formed in the neck step, the neck portion may be made to conform to the melt and the neck portion may be subsequently formed.
(2) When using the MCZ method, first, the raw material of the crystal in the crucible is melted, and the seed crystal is immersed in a melt held in the crucible to make the seed crystal conform. Then, the neck process which raises a seed crystal and forms a neck part is implemented. Next, a shoulder portion and a body portion of a single crystal are formed. In such a single crystal manufacturing method, after the neck portion having a predetermined length is formed in the neck step, the neck portion may be continuously formed by fitting the neck portion into the melt.
When the MCZ method is applied, the transverse magnetic field applied to the melt may be performed in the range of 2000G to 4000G.
(3) In the manufacture of the silicon single crystal of the above (1) or (2), the length of the neck portion formed first in the neck step may be 20 mm or more. Further, in the neck step, after raising the temperature of the melt at the time of forming the neck portion, the neck portion may be adapted to the melt. Further, the temperature of the heater for melting the crystal raw material in the crucible may be measured (temperature measurement), and the temperature of the melt may be adjusted by controlling the temperature of the heater based on the temperature measurement result.

The above-mentioned adjustment of the seed crystal to the melt (silicon melt) and the conformation of the neck portion to the melt include the meniscus shape of the contact interface when the crystal is brought into contact with the melt, for example, Includes observing the overhang of crystal habit lines. That is, the temperature of the surface of the melt is estimated, the heater power (electric power) is controlled based on the temperature, the amount of heat input to the melt is adjusted, and the temperature of the surface of the melt is stabilized.
As described above, when the seed crystal and the neck portion are adapted to the melt, dislocations are easily removed. However, even when the seed crystal and the neck portion are not adapted to the melt, in any of the above manufacturing methods in the present invention, it is possible to manufacture a straight body portion having no axial dislocation.

The silicon single crystal of the present invention is produced by any one of the production methods described above.
The silicon wafer of the present invention is manufactured from the above silicon single crystal.

The method for producing a silicon single crystal of the present invention includes the following steps in the method for producing a silicon single crystal in which the silicon single crystal is pulled up by the CZ method. A dipping process in which the seed crystal is brought into contact with the silicon melt to start pulling. A dislocation removing step for forming a dislocation removing portion that removes dislocations generated in the dip step by extending outwardly according to crystal pulling or by annihilating them in a loop shape. A diameter expansion process that raises the shoulder that expands to the required diameter. A straight body process that raises the straight body. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. A lifting condition setting step of determining a dislocation-free implementation raising condition based on the determination of dislocation-free in the non-dislocation determination step. This solved the above problem.
The method for producing a silicon single crystal of the present invention includes the following steps in the method for producing a silicon single crystal in which the silicon single crystal is pulled up by the CZ method. A dipping process in which the seed crystal is brought into contact with the silicon melt to start pulling. A dislocation removing step for forming a dislocation removing portion that removes dislocations generated in the dip step by extending outwardly according to crystal pulling or by annihilating them in a loop shape. A diameter expansion process that raises the shoulder that expands to the required diameter. A straight body process that raises the straight body. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information.
The diameter expansion step may be started based on the determination of dislocation-free in the dislocation-free determination step.
In the above-described silicon single crystal manufacturing method, the high-energy radiation light may have an energy of 40 keV to 70 keV, and may be irradiated to the dislocation removing unit as a rotation state at a rotation speed of 0.1 to 30 rpm.

The method for producing a silicon single crystal of the present invention includes the following steps in the method for producing a silicon single crystal in which the silicon single crystal is pulled up by the CZ method. A dipping process in which the seed crystal is brought into contact with the silicon melt to start pulling. A dislocation removing step for forming a dislocation removing portion that removes dislocations generated in the dip step by extending outwardly according to crystal pulling or by annihilating them in a loop shape. A diameter expansion process that raises the shoulder that expands to the required diameter. A straight body process that raises the straight body. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. A lifting condition setting step of determining a dislocation-free implementation raising condition based on the determination of dislocation-free in the non-dislocation determination step.
This makes it possible to clarify the behavior of dislocations whose behavior has not been clarified. These dislocations can be eliminated so as to escape outward as the dislocation removal portion (dislocation-free portion) grows, or can be eliminated as a loop. Thereby, a dislocation-free part can be made into the dislocation-free state which does not have an axial dislocation. And it becomes possible to obtain | require easily the pulling conditions at the time of enlarging (diameter expanding) the diameter of a silicon single crystal, after making a dislocation-free part dislocation-free.

  In the dislocation-free determination step, the dislocation state is determined from the information acquired in the previous step. Therefore, it is possible to accurately grasp the dislocation behavior inside the dislocation removal unit (dislocation-free unit). Therefore, it is possible to effectively use the removal of dislocations due to the occurrence of a loop that was not previously known. And since the diameter of a silicon single crystal is made small, the diameter of the dislocation-free part conventionally called the neck part can be made larger than before. Therefore, even if the weight of the crystal pulled from the silicon melt increases, the silicon single crystal can be supported by the dislocation-free portion and the silicon single crystal having a large diameter can be pulled. Therefore, it is possible to omit an extra procedure related to a dip process or the like that increases the manufacturing time and the work process. In addition, it is possible to pull up a large silicon single crystal having a diameter of about 450 mm in a dislocation-free state having no axial dislocation.

The method for producing a silicon single crystal of the present invention includes the following steps in the method for producing a silicon single crystal in which the silicon single crystal is pulled up by the CZ method. A dipping process in which the seed crystal is brought into contact with the silicon melt to start pulling. A dislocation removing step for forming a dislocation removing portion that removes dislocations generated in the dip step by extending outwardly according to crystal pulling or by annihilating them in a loop shape. A diameter expansion process that raises the shoulder that expands to the required diameter. A straight body process that raises the straight body. A dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of dislocations in the dislocation removal unit by irradiating the dislocation removal unit with high energy radiation. A dislocation-free determination step of determining that dislocations are removed based on the dislocation behavior information. And based on the determination of non-dislocation in the non-dislocation determination step, the diameter expansion step is started.
As a result, it becomes possible to observe in real time the behavior of dislocations that occurred during the pull-up by nondestructive inspection, which was impossible in the past. Therefore, after confirming that the dislocations in the dislocation removing part (dislocation-free part) are removed and dislocation-free, the silicon single crystal is pulled up to form the shoulder part (expanded part) and the straight body part. Can do. Therefore, dislocation-free crystals having no axial dislocations in the shoulder portion and the straight body portion can be pulled up safely and easily.

In the silicon single crystal manufacturing method, the high-energy radiation light may have an energy of 40 keV to 70 keV.
Thereby, high energy synchrotron radiation has sufficient energy to acquire information on dislocation states necessary for growth of dislocation-free crystals.
Further, the dislocation removing unit may be irradiated with the high energy radiation light as a rotation state at a rotation speed of 0.1 to 30 rpm.
Thereby, the three-dimensional information of dislocation can be acquired. In addition, non-destructive three-dimensional information of dislocations can be acquired.
Here, the irradiation with high energy radiation light in a rotating state includes the following. The dislocation removing unit (dislocation-free unit) rotates with the central axis in the direction of crystal growth as the rotation center axis, and the dislocation removing unit is irradiated with high energy radiation. High energy synchrotron radiation for acquiring dislocation information is performed in a relative motion state where the irradiation source of the high energy synchrotron radiation and the dislocation removal unit are comparable.

According to the present invention, it is possible to clarify the behavior of difficult-to-remove dislocations whose behavior has not been clarified. Further, this dislocation can be eliminated so as to escape outward along with the growth of the dislocation removing portion. Alternatively, dislocations can be eliminated as loops.
Thus, by eliminating the dislocations, the dislocation-free portion is made dislocation-free, the silicon single crystal is pulled up to increase the diameter, and the straight body portion can be grown. Alternatively, after confirming that the dislocation-free portion has been dislocation-free, it is possible to grow a silicon single crystal to grow a shoulder portion (expanded diameter portion) and a straight body portion. Thereby, the diameter required to support the weight of the silicon single crystal can be maintained. Therefore, dislocation-free crystals having no large weight of axial dislocations can be safely and easily pulled up from the silicon melt.

  According to the method for producing a silicon single crystal of the present invention, dislocations generated in the step of bringing the seed crystal into contact with the silicon melt can be removed, and dislocation-free can be realized. In addition, it is possible to accurately grasp the occurrence and removal of dislocations. In addition, it is possible to accurately grasp the relationship between the state of dislocation at the neck and the conditions for pulling up the silicon single crystal. In addition, the pulling conditions for the silicon single crystal that can be dislocation-free can be accurately determined.

It is sectional drawing which shows the outline of the manufacturing apparatus used for the manufacturing method of the silicon single crystal which concerns on 1st Embodiment. It is a flowchart of the manufacturing method of the silicon single crystal which concerns on 1st Embodiment. (A) to (e) are explanatory diagrams of a step S02 and a step S03 according to the first embodiment. It is explanatory drawing of process S12 which concerns on 1st Embodiment. (A) And (b) is a graph which shows the characteristic of the high energy radiation light in process S12 which concerns on 1st Embodiment. It is explanatory drawing of process S12 which concerns on 1st Embodiment. It is image data of the dislocation-free part acquired in process S12 concerning a 1st embodiment. It is image data of a dislocation-free part acquired in step S12 according to the first embodiment, (a) shows a state where dislocations are not removed, and (b) shows a state where dislocations disappear in a loop shape. (C) shows a state where dislocations are removed. It is sectional drawing which shows the outline of the manufacturing apparatus used for the manufacturing method of the silicon single crystal which concerns on 2nd Embodiment. It is a flowchart of the manufacturing method of the silicon single crystal which concerns on 2nd Embodiment. It is image data which shows the observation result of a wafer, (a) is an example of an observation of a wafer which has an axial dislocation, (b) is an example of an observation of a wafer which does not have an axial dislocation. It is the schematic of the silicon single crystal manufactured by the conventional dash necking method. It is an enlarged view of the silicon single crystal manufactured by the conventional dash necking method. It is sectional drawing which shows the outline of the manufacturing apparatus used for the manufacturing method of the silicon single crystal which concerns on 3rd Embodiment. It is sectional drawing which shows the outline of the manufacturing apparatus used for the manufacturing method of the silicon single crystal which concerns on 3rd Embodiment. (A) to (c) is an explanatory view illustrating step S03 according to the third embodiment. (A) And (b) is explanatory drawing explaining process S03 which concerns on 3rd Embodiment. It is image data of the dislocation-free part acquired in step S12 according to the third embodiment, (a) shows a state where dislocations are not removed, and (b) and (c) show dislocations removed. Indicates the state.

Hereinafter, a method for producing a silicon single crystal according to a first embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic front view showing an apparatus for producing a silicon single crystal in the present embodiment.
As shown in FIG. 1, a CZ furnace which is a silicon single crystal manufacturing apparatus using the Czochralski method (hereinafter referred to as CZ method) includes a crucible 1, a heater 2, a pulling shaft 4, a seed chuck 5, , A heat shield (heat shield) 7 and a magnetic field generator (magnetic field supply device) 9 are provided.

The crucible 1 is disposed in the center of the chamber and has a double structure having a quartz crucible 1a and a graphite crucible 1b. The quartz crucible 1a accommodates the silicon melt 3 inside. The graphite crucible 1b is disposed outside the quartz crucible 1a and holds the quartz crucible 1a. The crucible 1 is rotated and raised / lowered by a support shaft (pedestal) 1c.
The heater 2 is disposed outside the crucible 1. For example, a resistance heater can be used as the heater 2. The heater 2 is not limited to a resistance heater as long as the temperature of the silicon melt 3 accommodated in the crucible 1 is heated to a temperature equal to or higher than the melting point and can be maintained.
The pulling shaft 4 rotates at an arbitrary rotational speed around an axis parallel to the vertical direction. Further, the lifting shaft 4 moves at an arbitrary speed in an axial direction parallel to the vertical direction.
The seed chuck 5 is provided at the lower end of the pulling shaft 4 and holds the silicon seed crystal T.
When the silicon single crystal 6 is grown in the CZ furnace shown in FIG. 1, the pulling shaft 4 is lowered and the seed crystal T attached to the seed chuck 5 is immersed in the silicon melt 3. Thereafter, while rotating the crucible 1 and the pulling shaft 4, the pulling shaft 4 is raised to pull up the seed crystal T and the silicon single crystal 6.

The heat shielding portion 7 is provided above the crucible 1 so as to surround the side surface of the growing silicon single crystal 6 and to surround a portion above the liquid surface of the silicon melt 3. The heat shielding portion 7 has a cylindrical graphite outer shell, and the outer shell is filled with graphite felt. The heat shielding portion 7 is provided between the heater 2 and the silicon melt 3 surface and the side surface portion of the silicon single crystal 6 to block radiant heat. The inner surface of the heat shield 7 is inclined so that the inner diameter of the heat shield 7 gradually decreases from the upper end toward the lower end. The outer surface of the upper part of the heat shielding part 7 is inclined in the same manner as the inner surface. The outer surface of the lower part of the heat shielding part 7 is substantially parallel to the vertical direction. Thereby, the thickness of the lower part of the heat shielding part 7 is thicker toward the lower side closer to the silicon melt 3.
The width (thickness) W in the radial direction of the lower part of the heat shielding part 7 is, for example, about 50 mm. The inclination θ with respect to the vertical direction of the inner surface of the heat shielding portion 7 that is the inverted truncated cone surface is, for example, about 21 °. The height H1 from the liquid surface of the silicon melt 3 to the lower end of the heat shield 7 is, for example, in the range from about 10 mm to about 250 mm. The height H1 may be 50 mm or 100 mm, for example. Moreover, you may set the height H1 in each process mentioned later, respectively.

  The magnetic field generator 9 is disposed outside the heater 2. The magnetic field generated by the magnetic field generator 9 may be a horizontal magnetic field or a cusp magnetic field. The strength of the horizontal magnetic field may be in a range from 2000 G to 5000 G (from 0.2 T to 0.5 T). The strength of the horizontal magnetic field may be in the range of 3000G to 4000G (0.3T to 0.4T). A more preferable range of the strength of the horizontal magnetic field is 3000G to 3500G (0.30T to 0.35T). The height of the magnetic field center may be from 150 mm below the liquid surface of the silicon melt 3 to 100 mm above the liquid surface. A more preferable height of the magnetic field center is in a range from 75 mm below the liquid level to 50 mm above the liquid level. In some cases, the magnetic field generator 9 may not generate a magnetic field.

Next, a method for producing a silicon single crystal using the CZ furnace shown in FIG. 1 will be described.
When the silicon single crystal 6 is manufactured, the state of dislocations generated and the behavior of the dislocations are changed by changing the pulling conditions. Therefore, in the method for manufacturing a silicon single crystal of the present embodiment, the pulling conditions for the silicon single crystal 6 that can remove dislocations are set.

  As shown in FIG. 2, the manufacturing method of this embodiment includes a pulling condition setting step S00, a pulling preparation step S01, a dipping step S02, a dislocation removing step S03, a diameter expanding step S04, and a straight body step S05. , Tail process S06, measurement preparation process S11, dislocation behavior information acquisition process S12, and non-dislocation determination process S13.

First, in step S00, a pulling condition that can make the silicon single crystal 6 dislocation-free is set. Here, the pulling conditions include the following. Holding time of the seed crystal (seed) T before contact with the silicon melt 3. The contact time of the silicon seed crystal T with the silicon melt 3 and the temperature of the seed crystal T based on the contact time. The distance (height H1) between the heat shield 7 and the silicon melt 3. A radial width W of the lower end portion of the heat shield portion 7. The conditions of the atmosphere in the furnace. The rotational speed and pulling speed of the pulling shaft 4 when pulling up the silicon single crystal. The number of rotations of the crucible 1. The length in the pulling direction of a dislocation-free portion (neck portion, dislocation removal portion) N described later. Application state of magnetic field. The heating state or temperature distribution of the grown silicon single crystal. These pulling conditions can be set to arbitrary values in the initial state before the production of the silicon single crystal 6 is started.
Note that as the pulling conditions for the silicon single crystal 6 in step S00, data obtained by previously pulling the silicon single crystal 6 a plurality of times under different conditions may be used. In this case, the pulling conditions are selected based on the specifications of the silicon single crystal 6 to be manufactured.

Next, in step S01, high-purity silicon polycrystal is put into the crucible 1. The weight of the silicon polycrystal to be added is in the range from about 100 kg to about 400 kg. At this time, based on whether the product is p-type or n-type, a dopant for setting the product to a predetermined resistivity is added to the silicon melt 3 so as to have a predetermined concentration. Further, the concentration of carbon and nitrogen for setting or adjusting the resistivity or gettering ability of the product is adjusted. In some cases, the dopant may not be added.
In the present embodiment, 300 kg of silicon polycrystal is put into the crucible 1. The concentration of the dopant is adjusted so that, for example, the resistance value of the straight body portion 6b of the silicon single crystal 6 to be manufactured is 12 Ωcm.
Here, the atmosphere inside the CZ furnace is an inert gas, and the pressure of the inert gas is adjusted within a range from 1.3 kPa to 13.3 kPa (from 10 torr to 100 torr). In the present embodiment, the atmosphere inside the CZ furnace is, for example, Ar gas having a pressure of 50 Torr (6.666 kPa). The atmosphere inside the CZ furnace may contain hydrogen gas.
Further, a horizontal magnetic field of, for example, 3000 G (0.3 T) is generated by the magnetic field generator 9, and the silicon polycrystal is heated and melted by the heater 2. At this time, the height of the center of the magnetic field is set in a range from 75 mm below the liquid surface of the silicon melt 3 to 50 mm above the liquid surface. When applying the CZ method (MCZ method) using a magnetic field, the transverse magnetic field applied to the silicon melt 3 may be within a range from 2000 G to 4000 G.

  Next, in step S02, as shown in FIG. 3A, the seed crystal T attached to the seed chuck 5 is brought close to the silicon melt 3 and heated while being held for a predetermined time. Thereafter, the seed crystal T is lowered while being rotated, and is brought into contact with the silicon melt 3 as shown in FIG. And the lower end part of the seed crystal T is immersed in the silicon melt 3 (dip process). At this time, a dislocation t is generated at the lower end of the seed crystal T due to heat shock (see arrow A1). At this time, the respective temperatures are set so that the temperature state of the seed crystal T and the silicon melt 3 falls within a predetermined range. Specifically, the surface temperature of the silicon melt 3 is adjusted by adjusting the heating state of the heater 2 so that a meniscus having a predetermined shape is formed around the tip of the seed crystal T.

  Next, in step S03, as shown in FIG. 3C, the silicon single crystal 6 is pulled up while the crucible 1 and the pulling shaft 4 are rotated. At this time, the shape of the interface between the silicon single crystal 6 and the silicon melt 3 is formed in a concave shape (see arrow A2). The conditions set in step S00 are used as the conditions for pulling up the silicon single crystal 6. At this time, depending on the pulling conditions, the dislocations generated in the step S02 extend toward the outside in the radial direction of the silicon single crystal 6 and disappear, or disappear in pairs in a loop shape. Thereby, as shown in FIG. 3D and FIG. 3E, the dislocation t is removed, and a dislocation-free neck portion, that is, a dislocation-free portion N is formed (see arrow A3). .

  When the dislocation-free portion N from which the dislocation t is removed is formed, as shown in FIG. 3C, the interface between the silicon single crystal grown on the lower end portion of the seed crystal T and the silicon melt 3 The shape of the (meniscus) becomes a predetermined shape. The shape of the meniscus formed at the lower end of the seed crystal T or the silicon single crystal 6 varies depending on the heating conditions of the heater 2 and the pulling conditions such as the lifting speed of the pulling shaft 4. For example, the shape of the meniscus formed at the lower end of the seed crystal T or the silicon single crystal 6 can be adjusted by adjusting the temperature of the seed crystal T and the silicon melt 3 to a temperature within a predetermined range. .

  When the dislocation-free portion N having no dislocation t is formed, as shown in FIG. 3D, the dislocation t below the dislocation t generated in step S02 is the diameter of the silicon single crystal 6. It is formed to extend outward in the direction. In addition, crystal habit lines are formed on the outer periphery of the silicon single crystal 6 (see arrow A4). The portion inside the crystal habit line in the radial direction is in a dislocation-free state having no dislocations. Further, when the silicon single crystal 6 grows, as shown in FIG. 3E, the lower side crystal habit line is continuous (see arrow A5), and dislocation-free portions are continuously formed (see FIG. 3E). (See arrow A6). Further, a dislocation-free initial cone is formed below the dislocation-free portion (see arrow A7). As described above, the dislocation t extends toward the outside in the radial direction of the silicon single crystal 6 and disappears, or disappears in a loop shape. Thereby, the dislocation-free part N is formed.

  Next, in step S04, the diameter of the silicon single crystal 6 is linearly increased by controlling the pulling speed of the silicon single crystal 6, the rotation speed of the pulling shaft 4, the rotation speed of the crucible 1, and the heating conditions of the heater 2. Let Thereby, the diameter of the lower side of the dislocation-free portion N in which the dislocation-free portion of the silicon single crystal 6 shown in FIG. As shown in FIG. 1, a conical shoulder portion 6 a is formed on the lower side of the dislocation-free portion N of the silicon single crystal 6.

  Next, in step S05, the pulling shaft 4 is pulled up to a predetermined length while maintaining the diameter of the silicon single crystal 6 at, for example, 300 mm or 450 mm, which is the diameter of the silicon wafer to be the product. Thereby, as shown in FIG. 1, the cylindrical straight body part 6b with a constant diameter is formed in the silicon single crystal 6 (straight cylinder process).

  Next, in step S06, the diameter of the silicon single crystal 6 is linearly reduced by controlling the pulling speed of the silicon single crystal 6, the rotation speed of the pulling shaft 4, the rotation speed of the crucible 1, and the heating conditions of the heater 2. Let Thereby, a conical tail portion (not shown) opposite to the shoulder portion 6a is formed in the silicon single crystal 6. Thereafter, the silicon single crystal 6 is separated from the silicon melt 3 by separating the tail portion and the silicon melt 3.

  Next, in step S11, preparation for measuring the dislocation state of the dislocation-free part N is made. Specifically, first, the dislocation-free portion N is separated from the silicon single crystal 6. As shown in FIG. 4, the separated dislocation-free portion N is supported by the support portion R <b> 1 so that it can rotate about the central axis N <b> 4 of the silicon single crystal that is coaxial with the pulling shaft 4. Here, the support portion R1 is provided so that the direction of the central axis N4 that is the rotation axis can be changed.

  Next, in step S12, as shown in FIG. 4, the non-dislocation portion N is irradiated with white X-rays as high energy radiation B from the irradiation portion R, and the non-dislocation portion N is irradiated with white X-rays. X-rays are detected by a detection unit d including a CCD (dislocation behavior information acquisition step). The irradiation unit R includes an absorber, a shutter, and a diffractometer, and has a shutter / absorber that is used for controlling time resolution and taking measures against heat load. The detection unit d has a function as a diffractometer. The detection unit d has sufficient performance for detecting white X-ray topography.

  Conventionally, analysis of dislocations to be detected by X-ray topography has been performed according to the following procedure. First, the cross section to be analyzed is sliced to a thickness of about 1.5 mm. Then, the sliced surface to be analyzed is etched with mixed acid, and irradiated with X-rays to observe a transmission diagram. Therefore, only a transmission diagram in a cross section with a predetermined interval can be obtained. For example, images at intervals of 1.5 mm are connected to grasp a change in dislocation state accompanying crystal growth. Therefore, only dislocations propagating obliquely with respect to the pulling axis of the silicon single crystal could be observed.

  On the other hand, when white X-rays are used as the high-energy radiation described in Non-Patent Document 1, when the dislocation-free portion N is observed, dislocations that disappear in a loop shape can be measured. Thereby, it can be confirmed that dislocations exist in addition to the normal {111} plane. By using such high-energy white X-rays, it is possible to observe the behavior of dislocations inside single crystal silicon as a three-dimensional image without performing a destructive inspection such as slicing. As a result, the position information of dislocations and their properties can be observed three-dimensionally. As such an observation apparatus, BL28B of a large synchrotron radiation facility (Spring-8) of the High-intensity Photoscience Research Center can be used.

  White X-rays have a continuous spectrum and have an energy in the range from 30 keV to 1 MeV. The energy of the white X-ray may be in the range from 40 keV to 100 keV, or in the range from 50 keV to 60 keV. The wavelength of white X-rays may be in the range from 0.001 nm to 0.25 nm. Thereby, the state of dislocation in the dislocation-free portion N can be sufficiently observed.

FIG. 5A is a graph showing the distribution of the number of photons obtained through a 1 mm × 1 mm slit at a distance of 44 m from the X-ray light source of the irradiation unit R, with the vertical axis representing the number of photons and the horizontal axis representing energy. It is.
FIG. 5B is a graph showing the distribution of the number of photons obtained through a 1 mm × 1 mm slit at a distance of 44 m from the X-ray light source of the irradiation unit R, where the vertical axis represents the number of photons and the horizontal axis represents the wavelength. It is.
Here, the X-ray beam diameter is preferably in the range of 0.01 to 1 times the diameter of the dislocation-free portion N that is the object to be measured.

  In the present embodiment, the non-dislocation portion N is irradiated with white X-rays having an energy of 40 keV to 70 keV, which is high-energy radiation light B, while rotating at a rotational speed of 0.1 rpm to 30 rpm. Specifically, the non-dislocation portion N is irradiated with high energy radiation light while rotating the non-dislocation portion N around the central axis N4 in the crystal growth direction as the rotation axis. Alternatively, the high-energy radiation light is irradiated from the irradiation unit R to the non-dislocation unit N while rotating the irradiation unit R about the central axis N4. Both the irradiation part R and the dislocation-free part N may be rotated about the axis N4 as a rotation axis. And the three-dimensional information of the dislocation of the dislocation-free part N can be acquired nondestructively by, for example, Fourier transforming the detection result.

In step S12, as shown in FIG. 4, the irradiation angle ω of the high-energy radiation light B is set with respect to the axis N4 of the dislocation-free portion N of the silicon single crystal 6. Then, paying attention to one diffraction spot (for example, 004 diffraction spot), a dislocation-free portion N is rotated to construct a CT image by a topographic image. Thereby, the three-dimensional position information of the dislocation line is obtained. The three-dimensional behavior of dislocations in the dislocation-free part N obtained in this way is recorded as first information.
Further, dislocation properties such as Burgers vector are measured by white X-ray topography shown in FIG. In FIG. 6, the detection positions d1 to d7 correspond to the diffraction direction of white X-rays. Each detection position is, for example, a position where a Laue topograph is acquired in the crystal orientations [-1 1 1], [-1 1 3], [0 0 4], [1 -1 3], and [1 -1 1]. It corresponds. Here, -1 in the above crystal orientation means 1 with an overline.

  Next, in step S13, it is determined whether or not the dislocation is removed based on the first information including the three-dimensional behavior of the dislocation acquired in step S12. Specifically, it is determined whether or not dislocation has been removed from the image data of the dislocation-free portion N shown in FIG. That is, as shown in FIG. 8A, when the dislocation extends linearly along the central axis N4, it is determined that the dislocation has not been removed. As shown in FIG. 8B, dislocations are present, but when these are pair-annihilated in a loop shape, it is determined that the dislocation has been removed. As shown in FIG. 8C, when the dislocation cannot be confirmed, it is determined that the dislocation has been removed.

Next, the determination of dislocation-free in step S13 is fed back to step S00.
That is, when dislocations in the dislocation-free portion N are removed by the pulling conditions initially set in step S00, the pulling conditions are recorded in the database, and the pulling conditions are used in step S00. When dislocations in the dislocation-free portion N are not removed by the pulling conditions initially set in the step S00, the pulling conditions are recorded in the database so that they are not used in the step S00. A new pulling condition that can be dislocation free is set.

As described above, according to the method for manufacturing a silicon single crystal of the present embodiment, dislocations N are formed, so that dislocations generated in the step of bringing the seed crystal T into contact with the silicon melt 3 are generated by the silicon. The single crystal 6 can be removed at the shoulder portion 6a and the straight body portion 6b. Thereby, dislocation-free of the shoulder part 6a and the straight body part 6b of the silicon single crystal 6 can be realized.
In addition, the state of dislocation generation and removal can be accurately grasped by the steps S12 and S13.
Further, by repeating the steps S00 to S06 and the steps S11 to S13, it is possible to accurately determine the pulling conditions for the silicon single crystal that can be dislocation-free.
Further, dislocations in the dislocation-free portion N can be removed without making the neck portion of the silicon single crystal 6, that is, the diameter of the dislocation-free portion N unnecessarily small. Therefore, a large silicon single crystal 6 having a diameter of about 450 mm from which dislocations have been removed can be safely grown by a simple method without requiring any special treatment or apparatus.

Hereinafter, a method for producing a silicon single crystal according to a second embodiment of the present invention will be described with reference to the drawings.
In the present embodiment, parts corresponding to those of the first embodiment described above are denoted by the same reference numerals and description thereof is omitted.

As shown in FIG. 9, the CZ furnace of the present embodiment includes a crucible 1, a heater 2, a pulling shaft 4, a seed chuck 5, a heat shielding unit 7, a magnetic field generation unit 9, an irradiation unit R, It has a detector d.
The irradiation part R irradiates the high-energy radiation light B to the dislocation-free part N being pulled up and located in the chamber. The detection part d detects the diffracted light of the high energy radiation light B irradiated to the dislocation-free part N. The irradiation part R and the detection part d are provided so that the angle ω of the high-energy radiated light B with respect to the pulling axis 4 substantially coincident with the axis N4 can be adjusted.

  As shown in FIG. 10, the method for manufacturing a silicon single crystal according to the present embodiment includes steps S00 to S06 as in the first embodiment. However, the first implementation described above is that the steps S11 to S13 are not provided after the steps S00 to S06, and the dislocation behavior information acquisition step S22 and the non-dislocation determination step S23 are provided between the steps S03 and S04. It is different from the form.

  In step S22, similarly to step S12, the non-dislocation unit N is irradiated with the high energy radiation light B from the irradiation unit R, and the diffracted light is measured by the detection unit d. In the present embodiment, the irradiation of the high energy radiation light B and the detection of the diffracted light to the dislocation-free portion N are performed while pulling up the silicon single crystal inside the CZ furnace.

  In step S23, it is determined whether or not the dislocation of the dislocation-free portion N has been removed, as in step S13. Then, when the dislocation of the dislocation-free part N has been removed, the process S04 is started. Further, when the dislocation of the dislocation-free portion N has not been removed, the process returns to step S03, where the dislocation-free portion N is further grown, or the dislocation-free portion N is regrown by meltback. In addition, the pulling condition is fed back to step S00, and a pulling condition different from the fed pulling condition is set.

  According to the present embodiment, the state shown in FIGS. 3A to 3E is measured in real time. For this reason, after confirming that the dislocation t has been removed in the dislocation-free portion N, the process S04 can be started. Therefore, the silicon single crystal 6 in the state where the dislocation t is not removed in the shoulder portion 6a and the straight body portion 6b is not pulled up. Therefore, the time required for manufacturing the silicon single crystal 6 can be shortened. Moreover, the yield in pulling up the silicon single crystal 6 can be remarkably improved.

Next, a third embodiment of the present invention will be described with reference to the drawings.
In the present embodiment, a method for producing a silicon single crystal capable of removing axial dislocations (hardly removed dislocations) extending in the pulling direction of the silicon single crystal over the entire length of the silicon single crystal will be described.

  An axial dislocation is a dislocation that occurs at the neck portion and cannot be removed even if the diameter of the silicon single crystal is reduced by a conventional dash necking method or the like. Conventionally, axial dislocations have been detected as pits when a wafer is manufactured by slicing a straight body portion of a silicon single crystal.

Axial dislocations can be detected separately from pits by the following method.
First, on the surface of a silicon single crystal wafer, a strain inspection apparatus (SIRD: registered trademark) SirTec manufactured by JENA WAVE can be used to observe the internal stress distribution state by an optical inspection means. taking measurement.

  Specifically, the wafer is subjected to a saliency treatment in which the wafer is heated for 1 sec or more at a temperature rising rate within a range from 10 ° C./sec to 300 ° C./sec and at a temperature within a range from 900 ° C. to 1250 ° C. Then, the surface properties of the wafer including the manifested distortion are evaluated. In this manifestation process, a large temperature difference is generated in the surface of the wafer due to rapid heating, and thermal stress due to this is generated. Therefore, heating for a long time is not necessary, and heating for a short time of about 1 sec is sufficient.

  Axial dislocations exist through the front and back surfaces of the wafer. Therefore, distortion of the axial dislocation increases due to the thermal stress generated by the above-described revealing process. On the other hand, in the surface contamination inspection, dislocations that are detected as pits but are not long enough to penetrate the front and back surfaces of the wafer do not increase in distortion due to thermal stress generated by the revealing process. Specifically, a dislocation whose length in the thickness direction of the wafer is less than about 10% of the thickness of the wafer does not increase in distortion due to thermal stress generated by the revealing process.

  For this reason, axial dislocations are manifested by the above-described manifestation process, but other pits and the like are not manifested by this manifestation process. FIG. 11A shows an example of observation of a wafer having a manifested axial dislocation. FIG. 11B shows an observation example of a wafer having no axial dislocation. In addition, it is thought that the distortion observed at the edge of the wafer is generated by supporting the wafer during the revealing process.

  Axial dislocation occurs at the stage where the seed crystal is brought into contact with the silicon melt. And it exists continuously in the direction in which the silicon single crystal grows, that is, the neck portion, the shoulder portion, and the straight body portion. Axial dislocations appear in specific areas when a silicon single crystal is processed into a wafer. Axial dislocations cannot be removed by conventional dash necking methods.

As shown in FIG. 12, a silicon single crystal 60 manufactured by a conventional dash necking method has a seed crystal T0, a neck portion N0, a shoulder portion 60a, a straight body portion 60b, and a tail portion 60c.
When the vicinity of the seed crystal T0 is enlarged, as shown in FIG. 13, the heat shock dislocation Jn generated in the seed crystal T0 and continuing to the neck portion N0 due to the heat shock when the seed crystal T0 is brought into contact with the silicon melt. Will occur. When there is a mismatch in lattice constant between the seed crystal T0 and the neck portion N0, a misfit dislocation Jm occurs in the neck portion N0.

  Of the thermal shock dislocations Jn and misfit dislocations Jm, those grown in the growth direction of the silicon single crystal may become axial dislocations J. Conventionally, the axial dislocation J has been observed as follows. First, as shown in FIG. 12, the straight body portion 60b of the silicon single crystal 60 is thinly cut to obtain a plurality of wafers W having a {110} cross section. The axial dislocation J is observed in the 10 o'clock direction and the 4 o'clock direction when the notches of each wafer W are aligned with the 12 o'clock direction of the watch. More specifically, the axial dislocation J is the center of the silicon single crystal in the range from 120 ° to 135 ° and the range from 315 ° to 350 °, where the position of the notch provided in the {100} direction is 0 °. It is observed in a region J1 symmetric with respect to the axis. Further, the axial dislocation J is observed in a region that is rotated 90 ° and 45 ° around the central axis in a region that is symmetrical with respect to the central axis of the silicon single crystal. That is, the axial dislocation J extends in the growth direction (center axis direction) of the silicon single crystal in the region J1. In FIG. 12, only the region J1 in the range from 120 ° to 135 ° and the range from 315 ° to 350 ° is illustrated with the notch position being 0 °.

  Note that the axial dislocations J may disappear in a silicon single crystal grown longer than the normal length. However, when a silicon single crystal wafer of a normal length is observed, axial dislocations J exist throughout the entire length of the silicon single crystal, and the axial dislocations J penetrate the silicon single crystal in the central axis direction. There were many.

In the process of growing a silicon single crystal when pulling up the neck portion, shoulder portion and straight body portion, the direction in which the heat shock dislocation Jn and the misfit dislocation Jm propagate has the following characteristics.
First, the heat shock dislocation Jn and the misfit dislocation Jm usually move on the {111} plane, which is a silicon slip plane.
Secondly, the heat shock dislocation Jn and the misfit dislocation Jm are usually introduced into the {111} plane.
Thirdly, the heat shock dislocation Jn and the misfit dislocation Jm rarely are introduced into the {111} plane and may move to other planes, but immediately move to the {111} plane.
Fourth, most of the heat shock dislocations Jn and misfit dislocations Jm occur when the seed crystal is brought into contact with the silicon melt.

  When the silicon single crystal for producing a <100> wafer is pulled up, the heat shock dislocation Jn and the misfit dislocation Jm move along the {111} plane which is the sliding surface of the silicon single crystal in the step of forming the neck portion. It is considered that the surface of the silicon single crystal is reached. However, according to the observation result by the above X-ray topography, regardless of the diameter of the neck portion, the initial heat shock dislocation Jn and misfit dislocation Jm are not the slip surface {111} plane, but the silicon single crystal and silicon. Extends in a direction perpendicular to the interface with the melt. Thereafter, the heat shock dislocation Jn and the misfit dislocation Jm may extend in the growth direction of the silicon single crystal without moving to the slip plane.

In this case, the heat shock dislocation Jn and the misfit dislocation Jm do not reach the surface of the silicon single crystal even when the diameter of the silicon single crystal increases to the product diameter, and are axial dislocations existing inside the straight body portion. May be. Thus, a silicon single crystal having axial dislocations in the straight body portion has a habit line similar to a silicon single crystal having no axial dislocations in the straight body portion, and is a non-defective product in an ingot state. It may not be possible to distinguish. The axial dislocation is confirmed as a defect when a wafer is manufactured from the ingot.
On the other hand, the heat shock dislocation Jn and the misfit dislocation Jm extending in the growth direction of the silicon single crystal may move to the {111} plane during the formation of the shoulder portion or the straight body portion.

  If the heat shock dislocation Jn and the misfit dislocation Jm can be moved to the {111} plane in the step of forming the neck portion, the diameter of the neck portion is not reduced and the seed crystal is not heated to a high temperature. Even so, axial dislocations can be removed. Therefore, if the heat shock dislocation Jn and the misfit dislocation Jm can be moved to the {111} plane in the neck portion, the product is not affected by the axial dislocation even if the shoulder portion and the straight body portion are formed. In addition, the number and time of remelting at the initial stage of the growth of the silicon single crystal can be suppressed. That is, if the heat shock dislocations Jn and misfit dislocations Jm can be moved to the {111} plane at the neck, a heavy silicon single crystal with no axial dislocations in the shoulder and the straight body can be grown at low cost. can do.

  As described above, the heat shock dislocation Jn and the misfit dislocation Jm at the time of forming the neck portion progress not perpendicular to the {111} plane but perpendicular to the interface between the silicon single crystal and the silicon melt. Therefore, in the present embodiment, a method for moving the heat shock dislocation Jn and the misfit dislocation Jm that do not exist in the {111} plane to the {111} plane in the step of forming the neck portion will be described. Specifically, the shape of the interface between the silicon single crystal and the silicon melt when the neck portion is formed is controlled. And X-ray topography is used in order to confirm the conditions.

Hereinafter, the manufacturing method of the silicon single crystal of this embodiment will be described in detail with reference to the drawings.
The method for manufacturing a silicon single crystal according to this embodiment includes steps S01 to S06 and steps S11 to S13, as in the method for manufacturing a silicon single crystal according to the first embodiment shown in FIG. Or similarly to the manufacturing method of the silicon single crystal in 2nd Embodiment shown in FIG. 10, it has process S01 to process S06, process S22, and process S23. The present embodiment is different from the first and second embodiments in that the axial dislocation is removed in step S03.

  Further, the CZ furnace of the present embodiment is different from the CZ furnace described in the first embodiment or the second embodiment in that it has a temperature adjusting mechanism for the dislocation-free portion N. Other configurations are the same as those of the CZ furnace of the first embodiment or the second embodiment.

As shown in FIG. 14, the CZ furnace of this embodiment includes a laser irradiation unit La and a gas supply unit G as a temperature adjustment mechanism of the dislocation-free portion N of the silicon single crystal. The laser irradiation unit La and the gas supply unit G are provided so as to rotate around the pulling shaft 4 in synchronization with the rotation of the pulling shaft 4. That is, the laser irradiation unit La and the gas supply unit G heat and cool the silicon single crystal rotating together with the pulling shaft 4 from a predetermined direction.
The laser irradiation unit La and the gas supply unit G may not be provided so as to rotate around the pulling shaft 4. In this case, the laser irradiation unit La and the gas supply unit G may be fixed, and the laser beam and the cooling gas may be intermittently irradiated and supplied from the laser irradiation unit La and the gas supply unit G at a predetermined timing.

  Alternatively, in the CZ furnace of this embodiment, as shown in FIG. 15, a notch 7 </ b> A that can be opened and closed is formed at the lower end portion of the heat shield portion 7 as a temperature adjustment mechanism of the dislocation-free portion N of the silicon single crystal. In this case, the cylindrical portion 7a is provided on the upper portion of the heat shielding portion 7, and the lid portion 7b having an opening through which the lifting shaft 4 is passed is provided at the upper end of the cylindrical portion 7a. A controller 41 that controls the raising and lowering of the lifting shaft 4 is provided on the lid 7b. The heat shield part 7, the cylindrical part 7 a, the lid part 7 b, the control part 41, and the lifting shaft 4 rotate integrally around the axis that coincides with the lifting shaft 4 in synchronization with the rotation of the lifting shaft 4. That is, the heat shield 7 does not shield the heat of the heater 2 only in a predetermined direction of the silicon single crystal that rotates with the pulling shaft 4. In other words, the heater 2 heats the silicon single crystal rotating together with the pulling shaft 4 from the predetermined direction via the notch 7A.

In the present embodiment, in step S02, first, as shown in FIGS. 3A and 3B, the seed crystal T is brought into contact with the silicon melt 3.
Thereafter, in step S03, the pulling shaft 4 is raised, and a silicon single crystal is grown below the seed crystal T to form a dislocation-free portion N.
In step S02, the seed crystal T is lowered while being rotated, and the lower end portion of the seed crystal T is immersed in the silicon melt 3. After the lower end portion of the seed crystal T is immersed in the silicon melt 3, the descent of the seed crystal T is stopped, and the seed crystal T and the silicon melt 3 are sufficiently blended. Then, a part of the lower end portion of the seed crystal T is dissolved in the silicon melt 3.

If the silicon melt 3 is in a stable state, when the seed crystal T is brought into contact with the silicon melt 3 in step S02, the surface tension of the silicon melt 3 is present at the interface between the seed crystal T and the silicon melt 3. Acts and a meniscus is formed.
However, immediately after the silicon polycrystal as a raw material is melted, the temperature of the silicon melt 3 is largely unstable due to local temperature fluctuations. For this reason, step S02 is performed after a predetermined time has elapsed after melting the silicon polycrystal as a raw material.

When the temperature at the liquid surface of the silicon melt 3 is too high, the lower end portion of the seed crystal T is melted and separated when the silicon single crystal is pulled up.
When the temperature at the liquid surface of the silicon melt 3 is too low, the crystal grows on the side surface of the lower end portion of the seed crystal T and is in an overhanging state. When the process shifts from step S02 to step S03 in such a state, a new dislocation is generated in the non-dislocation portion N.

  When shifting from step S02 to step S03, it is necessary that the temperature of the silicon melt 3 is stable. Therefore, after immersing the lower end portion of the seed crystal T in the silicon melt 3, the lower end portion of the seed crystal T is sufficiently adjusted to the silicon melt 3, and the temperature of the silicon melt 3 is stabilized. The process proceeds to S03.

  Therefore, when the lower end portion of the seed crystal T is adapted to the silicon melt 3, the shape of the interface between the seed crystal T or the silicon single crystal and the silicon melt 3 is immersed after the seed crystal T is immersed in the silicon melt 3. Observe. Also, observe the interface to observe the overhang of crystal habit lines. Then, the temperature of the liquid surface of the silicon melt 3 is estimated from the shape of the interface between the seed crystal T or the silicon single crystal and the silicon melt 3 and the protrusion of the crystal habit line. And the electric power supplied to the heater 2 is controlled based on the estimated temperature, and the heat input to the silicon melt 3 is adjusted. That is, when the lower end portion of the seed crystal T is adapted to the silicon melt 3, a meniscus having a predetermined shape is formed at the lower end portion of the seed crystal T in a state where the growth rate of the seed crystal T is zero. The heating conditions by the heater 2 are adjusted. And the temperature of the liquid surface of the silicon melt 3 is adjusted, and the silicon melt 3 is stabilized.

  As described above, after the seed crystal T is adapted to the silicon melt 3 in the step S02, the silicon single crystal is pulled up to a predetermined length in the step S03 to form the dislocation-free portion N with a predetermined length. The length of the dislocation-free portion N formed after the seed crystal T is made to conform to the silicon melt 3 is desirably 20 mm or more. Thereafter, the lower end portion of the formed dislocation-free portion N having a predetermined length is made to conform to the silicon melt 3 in the same manner as the seed crystal T, and then the dislocation-free portion N is formed to a predetermined length.

  In the present embodiment, in step S03, the lower end portion of the seed crystal T is adjusted to the silicon melt 3 in a state where the temperature of the silicon melt 3 is controlled as described above. Forming N. And after making the lower end part of the neck part of the predetermined length formed into the silicon melt 3 in the state where temperature control was carried out similarly to the lower end part of the seed crystal T, the dislocation-free part N having a predetermined length was further added. Form. Thereby, the length of the dislocation-free portion N necessary for dislocation-freeness can be shortened, and dislocations and axial dislocations in the dislocation-free portion N can be reliably removed.

  Further, in step S03, after the temperature of the dislocation-free portion N is raised to a temperature higher than the temperature of the liquid surface of the silicon melt 3, the dislocation-free portion N may be adapted to the silicon melt 3. . Furthermore, the temperature of the heater 2 at the time of melting the silicon polycrystal as a raw material may be measured, and the temperature of the heater 2 may be controlled based on the measurement result to adjust the temperature of the silicon melt.

  The silicon single crystal grows substantially perpendicular to the interface K with the silicon melt 3 that is the surface on which the crystal is grown. Therefore, as shown in FIG. 16A, when the interface K between the silicon single crystal and the silicon melt 3 is flat, the dislocation j progresses to the {111} plane that is the sliding surface of the silicon single crystal. Instead, it develops in a direction substantially perpendicular to the interface K (the arrow direction in the figure). Therefore, when the interface K is a substantially horizontal flat surface, the dislocation j may be an axial dislocation extending from the interface K to the shoulder portion and the straight body portion of the silicon single crystal in the growth direction of the silicon single crystal.

  The shape of the interface between the silicon single crystal and the silicon melt 3 varies depending on the pulling speed of the silicon single crystal, that is, the lifting speed of the pulling shaft 4. Therefore, it is considered that the rising speed of the lifting shaft 4 is changed to change the shape of the flat interface K substantially parallel to the horizontal plane shown in FIG.

When the pulling speed of the silicon single crystal is lower than a predetermined speed, the interface K1 between the silicon single crystal and the silicon melt 3 changes to a downwardly convex shape as shown in FIG. When the pulling speed of the silicon single crystal is higher than a predetermined speed, the interface K2 between the silicon single crystal and the silicon melt 3 changes to an upwardly convex shape as shown in FIG.
That is, by controlling the pulling rate of the silicon single crystal, the shapes of the interfaces K1, K2 between the silicon single crystal and the silicon melt 3 can be controlled.

The silicon single crystal grows in a direction perpendicular to the interfaces K1 and K2. Further, the dislocation j progresses in the direction in which the silicon single crystal grows. That is, the dislocation j progresses in the normal direction of a virtual tangent plane between the interfaces K1 and K2 between the silicon single crystal and the silicon melt 3.
Therefore, as shown in FIG. 16B, when the interface K1 is a downwardly convex curved surface, most of the dislocations j progress outward in the radial direction of the dislocation-free portion N of the silicon single crystal. As shown in FIG. 16C, when the interface K2 is an upwardly convex curved surface, most of the dislocations j progress toward the radially inner side of the dislocation-free portion N of the silicon single crystal.

  In this way, by controlling the shape of the interface K and deforming it into a curved surface, the direction in which the dislocation j advances can be moved in a direction different from the direction of the central axis, which is the growth direction of the silicon single crystal. . That is, at an arbitrary point on the interfaces K1 and K2, the dislocation j advances in the normal direction of the virtual tangent plane at a point where the virtual tangent plane is not horizontal. Therefore, at an arbitrary point on the interface K1, K2, the direction in which the dislocation j advances is different from the direction of the central axis, which is the growth direction of the silicon single crystal, in that the virtual tangent plane is not horizontal. To do. Thereby, the heat shock dislocation Jn and the misfit dislocation Jm that do not exist in the {111} plane can be moved to the {111} plane and removed.

That is, when the interface K1 is a downwardly convex curved surface, the dislocation j advances toward the outer side in the radial direction of the dislocation-free portion N of the silicon single crystal and reaches the outer peripheral surface of the silicon single crystal, thereby dislocation-free. Part N is removed. In addition, when the interface K2 is an upwardly convex curved surface, the dislocation j progresses inward in the radial direction of the dislocation-free portion N of the silicon single crystal and is eliminated at the dislocation-free portion N by annihilating the pair. The
Therefore, the dislocation j extending from an arbitrary point on the interfaces K1, K2 whose virtual tangent plane is not horizontal can be removed, and the occurrence of the axial dislocation J can be prevented.

When dislocation j is advanced toward the central axis of the silicon single crystal to be annihilated, or is advanced toward the outer peripheral surface of the silicon single crystal and removed, the dislocation j is formed when the pulling speed of the silicon single crystal fluctuates. The diameter of the silicon single crystal may vary. At this time, even if the diameter of the silicon single crystal is reduced to some extent, there is no problem as long as the diameter of the silicon single crystal finally formed can be held. Even if the diameter of the silicon single crystal is increased to some extent, there is no problem as long as the raw material is not wasted. Therefore, the diameter of the silicon single crystal may vary to some extent as long as the pulling of the straight body portion of the silicon single crystal is not adversely affected.
Due to the change in pulling speed, the direction in which dislocations progress may deviate from the pulling direction of the silicon single crystal. This phenomenon is preferable for dislocation removal.

  Interfaces K1 and K2 shown in FIGS. 16A and 16B are curved surfaces with a central axis parallel to the growth direction of the silicon single crystal. That is, the interfaces K1 and K2 are rotationally symmetric curved surfaces when the central axis of the silicon single crystal is the rotation axis. Therefore, the imaginary tangent plane at the intersection of the interfaces K1, K2 and the central axis of the silicon single crystal may be horizontal. In this case, the dislocation j generated on the central axis of the silicon single crystal advances in the growth direction of the silicon single crystal on the central axis of the silicon single crystal. This dislocation j may be an axial dislocation extending from the interfaces K1 and K2 to the shoulder portion and the straight body portion of the silicon single crystal.

  Therefore, in step S03 of the present embodiment, as shown in FIGS. 17A and 17B, the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 are curved, and silicon It is deformed into an asymmetric shape with respect to the central axis of the single crystal. Specifically, a temperature distribution in the direction perpendicular to the central axis of the silicon single crystal, that is, the radial direction is generated at the interfaces K3 and K4.

When a temperature distribution in the radial direction, that is, a temperature gradient in the radial direction does not occur in the silicon single crystal, as shown in FIGS. 16B and 16C, the dislocation-free portion N and the silicon melt 3 The interfaces K1 and K2 are curved surfaces symmetrical with respect to the central axis parallel to the growth direction of the silicon single crystal.
Here, by using a CZ furnace as shown in FIG. 14 or FIG. 15, the dislocation-free portion N of the silicon single crystal is heated or cooled from one direction, so that the interface K1, K2 has the center axis direction of the silicon single crystal. A temperature distribution perpendicular to is generated.

When the CZ furnace shown in FIG. 14 is used, the vicinity of the interfaces K1, K2 of the dislocation-free portion N is heated from one direction by irradiating laser light from the laser irradiation portion La.
Further, the cooling gas is supplied from the side opposite to the laser irradiation unit La by the gas supply unit G. Thereby, the vicinity of the interfaces K1, K2 of the dislocation-free portion N is cooled from the direction opposite to the direction in which the dislocation-free portion N is heated.

When the CZ furnace shown in FIG. 15 is used, the radiant heat of the heater 2 and the silicon melt 3 reaches the vicinity of the interfaces K1 and K2 of the non-dislocation portion N from the notch 7A of the heat shielding portion 7 so that no dislocation occurs. Part N is heated from one direction.
In this way, a temperature gradient perpendicular to the rising direction of the lifting shaft 4 is generated at the interfaces K1 and K2.

  Then, as shown in FIGS. 17A and 17B, the interfaces K3 and K4 between the dislocation-free portion N of the silicon single crystal and the silicon melt 3 are parallel to the growth direction of the silicon single crystal. Asymmetrical curved surface with a central axis. As a result, the horizontal portion Kd of the interfaces K3 and K4 moves to a position away from the portion Kc on the central axis of the silicon single crystal outward in the radial direction.

  The shapes of the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 are such that the strength of the magnetic field generated by the magnetic field generator 9, the rotational speed of the crucible 1, and the rotational speed of the silicon single crystal, that is, the pulling shaft 4 It also changes depending on the rotation speed. Therefore, by controlling these, the shapes of the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 are more precisely controlled, and the position of the horizontal portion Kd is more precisely controlled.

For example, the rotational speed of the crucible 1 and the rotational speed of the pulling shaft 4 may be controlled so that the silicon melt 3 and the silicon single crystal are relatively stationary. In this way, when the silicon melt 3 and the silicon single crystal are relatively stationary, fluctuation occurs in the convection of the silicon melt 3. Thereby, the convection of the silicon melt 3 is not symmetric with respect to the central axis of the silicon single crystal, and the growth of the silicon single crystal is not symmetric with respect to the central axis of the silicon single crystal. That is, the temperature distribution at the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is not symmetrical with respect to the central axis of the silicon single crystal. As a result, the horizontal portion Kd of the interfaces K3 and K4 moves to a position away from the portion Kc on the central axis of the silicon single crystal outward in the radial direction.

  Further, the strength of the magnetic field generated by the magnetic field generation unit 9 may be reduced to make the strength of the magnetic field zero. The convection of the silicon melt 3 is rectified by a magnetic field. For this reason, if the magnetic field is extinguished during the pulling of the silicon single crystal, the convection of the silicon melt 3 fluctuates. Thereby, the convection of the silicon melt 3 is not symmetric with respect to the central axis of the silicon single crystal, and the growth of the silicon single crystal is not symmetric with respect to the central axis of the silicon single crystal. That is, the temperature distribution at the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is not symmetrical with respect to the central axis of the silicon single crystal. As a result, the horizontal portion Kd of the interfaces K3 and K4 moves to a position away from the portion Kc on the central axis of the silicon single crystal outward in the radial direction.

  Further, the heater 2 may be rotated in synchronization with the rotation of the lifting shaft 4 in a state where the heater 2 is disposed asymmetrically on the lifting shaft 4. As a result, the convection of the silicon melt 3 is not symmetrical with respect to the central axis of the silicon single crystal, and the temperature distribution of the silicon melt 3 is not subject to the central axis of the silicon single crystal. Thereby, the growth of the silicon single crystal is not symmetrical with respect to the central axis of the silicon single crystal. That is, the temperature distribution at the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is not symmetrical with respect to the central axis of the silicon single crystal. As a result, the horizontal portion Kd of the interfaces K3 and K4 moves to a position away from the portion Kc on the central axis of the silicon single crystal outward in the radial direction.

  Further, the properties of a part of the inner peripheral surface of the quartz crucible 1a may be different from the properties of other parts. Specifically, a portion of about one quarter of the entire circumference of the inner peripheral surface within a range of about 10 cm from the upper end of the quartz crucible 1a may be a modified portion different from the other portions. In this case, in the altered portion, the bubble content at a position where the depth from the surface is within the range of 0.5 mm to 1 mm is within the range of 25% to 35% than the bubble content of other portions. Do more. For example, the bubble content rate in the altered portion may be increased by 30% from the bubble content rate in other portions.

  Thus, by making the properties of a part of the inner peripheral surface of the quartz crucible 1a different from the properties of the other parts, the convection of the silicon melt 3 is not symmetrical with respect to the central axis of the silicon single crystal, and the silicon single crystal Crystal growth is no longer symmetrical about the central axis of the silicon single crystal. That is, the temperature distribution at the interfaces K3 and K4 between the silicon single crystal and the silicon melt 3 is not symmetrical with respect to the central axis of the silicon single crystal. As a result, the horizontal portion Kd of the interfaces K3 and K4 moves to a position away from the portion Kc on the central axis of the silicon single crystal outward in the radial direction.

  In addition to the above method, by changing parameters such as the heating state of the heater 2, the strength of the magnetic field generated by the magnetic field generator 9, the rotational speed of the crucible 1, and the rotational speed of the pulling shaft 4, the interface K3 Any method that moves the position of the horizontal portion Kd of K4 can be similarly employed.

  In this way, by moving the horizontal portion Kd of the interfaces K3 and K4 from the central axis of the silicon single crystal to the outside in the radial direction, the direction in which the dislocation j generated on the central axis of the silicon single crystal propagates is increased. The direction is different from the direction of the central axis, which is the growth direction of the silicon single crystal. Further, the horizontal portion Kd of the interfaces K3 and K4 can be moved as appropriate. Thereby, the heat shock dislocation Jn and the misfit dislocation Jm that do not exist in the {111} plane can be reliably moved to the {111} plane and removed in the dislocation-free portion N. Therefore, the dislocation j extending from any point on the interfaces K3 and K4 can be reliably removed, and the axial dislocation J can be more reliably removed.

  Therefore, according to the present embodiment, in addition to the effects in the silicon single crystal manufacturing method described in the first and second embodiments, the axial dislocation extending to the straight body portion that could not be removed by the conventional method. It is possible to produce a silicon single crystal that does not exist. A wafer having no axial dislocations can be manufactured from the straight body portion of the silicon single crystal in which no axial dislocations exist.

  As mentioned above, although preferable embodiment of this invention was described, this invention is not limited to said embodiment. Additions, omissions, substitutions, and other modifications can be made without departing from the spirit of the present invention. The present invention is not limited by the above description, but only by the scope of the appended claims.

In this embodiment, 300 kg of silicon polycrystal is put into a crucible. The concentration of the dopant is adjusted so that the resistance value of the straight body portion of the silicon single crystal to be manufactured is 12 Ωcm. The atmosphere inside the CZ furnace is Ar gas having a pressure of 50 Torr (6.666 kPa).
The dislocation-free part is grown with the diameter of the seed crystal being 300 mm.
Here, using the five different conditions from Sample 1 to Sample 5 shown in Table 2 as the pulling conditions in Step S00 shown in FIG. 2, Steps S00 to S06 and Steps S11 to S13 were performed. Specifically, as shown in Table 2, the pulling speed of the pulling shaft, the distance (height) from the heat shield and the liquid surface of the silicon melt, and the seed crystal are immersed and held in the silicon melt. Five combinations with different retention times were used.

In step S13 shown in FIG. 2, the state of each sample was observed. In sample 1, as shown in FIG. 8A, dislocations extend in the axial direction, and dislocations are not removed. In Sample 4, as shown in FIG. 8B, dislocations are present, but the dislocations are annihilated in a loop shape, and the dislocations are removed. In sample 5, dislocations are removed as shown in FIG.
Therefore, as the pulling conditions in step S00, the pulling speed of the pulling shaft is in the range from 0.5 mm / min to 1.5 mm / min, and the distance (height) from the silicon melt is in the range of 100 mm to 150 mm. The holding time is set to 100 min. As a result, dislocations at the dislocation-free portion were removed, and dislocations at the shoulder portion and the straight body portion of the silicon single crystal could be removed.

When a silicon single crystal is manufactured using a conventional method, an axial dislocation extending in the central axis direction of the silicon single crystal exists at the neck portion of the silicon single crystal, as shown in FIG. Yes.
Next, in the same manner as in Example 1, axial dislocations were removed.
As a result, in step S13 shown in FIG. 2, as shown in FIG. 18B, it is determined that the axial dislocation has been removed in the dislocation-free portion. Alternatively, as shown in FIG. 18 (c), the dislocation progress direction moves to a {1 1 1} plane different from the direction of the central axis of the silicon single crystal, and the dislocation is dislocation-free at the silicon single crystal. It reaches a state where it has been removed by reaching the outer peripheral surface of.

  The method for producing a silicon single crystal of the present invention includes the following steps. A step of growing a silicon single crystal by bringing the seed crystal into contact with the silicon melt and then pulling it up. A step of forming dislocation-free portions in the silicon single crystal by causing dislocations generated in the silicon single crystal to be removed by extending outward in the radial direction of the silicon single crystal or eliminating them in a loop; A step of pulling up the silicon single crystal on which the dislocation-free portion is formed and expanding it to a predetermined diameter to form a straight body portion; The method for producing a silicon single crystal of the present invention solves at least one of the following problems. Dislocations generated in the process of bringing the seed crystal into contact with the silicon melt are removed to realize dislocation-free. To accurately grasp the occurrence and removal of dislocations. To accurately grasp the relationship between the state of dislocations at the neck and the conditions for pulling up the silicon single crystal. To accurately determine the pulling conditions for silicon single crystals that can be dislocation-free.

1a quartz crucible, 3 silicon melt, 4 pulling axis (center axis), 6 silicon single crystal, 6a shoulder part, 6b straight body part, B high energy radiation, N dislocation-free part (dislocation removal part, difficult to remove dislocation) Removal part), T seed crystal, K1, K2, K3, K4 interface (solid-liquid interface)

Claims (6)

In the method for producing a silicon single crystal in which the silicon single crystal is pulled by the CZ method,
A dipping step of bringing a seed crystal into contact with the silicon melt and starting to pull up the silicon single crystal;
Dislocations generated in the dipping process are removed by extending the silicon single crystal in the outward direction according to the pulling up of the silicon single crystal or by annihilating it in a loop form, and dislocation removal or non-dislocation parts in the silicon single crystal Dislocation removal step to form
A diameter expanding step of pulling up the silicon single crystal so as to form a shoulder portion by expanding the diameter of the silicon single crystal to a required diameter;
A straight cylinder step of pulling up the silicon single crystal so as to form a straight cylinder portion having the required diameter in the silicon single crystal;
Dislocation behavior information acquisition step of detecting dislocation behavior information which is a three-dimensional behavior of the dislocation in the dislocation removal unit or the non-dislocation unit by irradiating the dislocation removal unit or the non-dislocation unit with high energy radiation. When,
Based on the dislocation behavior information, a dislocation-free determination step for determining that the dislocation has been removed,
Based on the determination of dislocation-free in the dislocation-free determination step, a pulling condition setting step for determining pulling conditions for dislocation-free the silicon single crystal;
I have a,
Wherein the high-energy synchrotron radiation, a method of manufacturing a silicon single crystal that will be irradiated with an energy in the range of 40keV~70keV in the transposition behavior information acquisition step. In the method for producing a silicon single crystal in which the silicon single crystal is pulled by the CZ method,
A dipping step of bringing a seed crystal into contact with the silicon melt and starting to pull up the silicon single crystal;
A straight body step of pulling up the silicon single crystal so as to form a straight body portion of the silicon single crystal to be a silicon wafer;
A dislocation removal or dislocation removal step that removes dislocations or difficult dislocations generated in the dip step before the straight body step; and
A diameter expanding step of pulling up the silicon single crystal so as to form a shoulder portion by expanding the diameter of the silicon single crystal to the diameter of the straight body portion after the difficult removal dislocation removing step;
Dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of the dislocation in the dislocation removal unit or the non-dislocation unit by irradiating the dislocation removal unit or the non-dislocation unit with high energy radiation;
Based on the dislocation behavior information, a dislocation-free determination step for determining that the dislocation has been removed,
Based on the determination of dislocation-free in the dislocation-free determination step, a pulling condition setting step for determining pulling conditions for dislocation-free the silicon single crystal;
And having
The high energy synchrotron radiation in the dislocation behavior information acquisition step is irradiated with energy in the range of 40 keV to 70 keV,
In the difficult-to-remove dislocation removal step, the dislocation generated in the dip step is removed by extending the silicon single crystal toward the outside according to the pulling up of the silicon single crystal or by annihilating it in a loop shape to form the silicon single crystal. A method for producing a silicon single crystal, which forms a dislocation removal part or a dislocation-free part.   The method for producing a silicon single crystal according to claim 2, wherein the diameter expansion step is started based on the determination of dislocation-free in the dislocation-free determination step. In the method for producing a silicon single crystal in which the silicon single crystal is pulled by the CZ method,
A dipping step of bringing a seed crystal into contact with the silicon melt and starting to pull up the silicon single crystal;
The dislocations generated in the dip process are removed by extending the silicon single crystal toward the outside according to the pulling of the silicon single crystal or by annihilating it in a loop shape, and dislocation removal or non-dislocation parts are formed in the silicon single crystal. Dislocation removal step to be formed;
A diameter expansion step of pulling up the silicon single crystal so as to form a shoulder portion by expanding the diameter of the silicon single crystal to a required diameter;
A straight cylinder step of pulling up the silicon single crystal so as to form a straight cylinder portion having the required diameter in the silicon single crystal;
Dislocation behavior information acquisition step of detecting dislocation behavior information that is a three-dimensional behavior of the dislocation in the dislocation removal unit or the non-dislocation unit by irradiating the dislocation removal unit or the non-dislocation unit with high energy radiation;
A non-dislocation determination step for determining that the dislocation has been removed based on the dislocation behavior information,
The high energy synchrotron radiation in the dislocation behavior information acquisition step is irradiated with energy in the range of 40 keV to 70 keV,
A method for producing a silicon single crystal, which starts the diameter expansion step based on the determination of dislocation-free in the dislocation-free determination step. In the silicon single crystal manufacturing method according to any one of claims 1, 2, 3, and 4 ,
The high energy synchrotron radiation is 0 . A method for producing a silicon single crystal irradiated at a rotational speed in a range of 1 rpm to 30 rpm. 6. The silicon single crystal according to claim 5 , wherein an irradiation region of an outer peripheral surface of the dislocation removal unit or the non-dislocation unit irradiated with the high energy light moves so as to go around the outer peripheral surface in the rotation state. Production method.